Department of Humanities, University of Maine at Farmington, Farmington, ME 04938, USA
email@example.com Submitted on 10 March 2003; Revised and Accepted on 28 Jan 2006
The present study compares oral and written narratives in Chinese and English, and demonstrates that a discourse of the same genre, same context, and same goal orientation, when produced in different medium, exhibits significant patterns of differences at various levels of structure between the modalities across languages. At the discourse level, the oral narrative is easily differentiated from the written one by speakers’ active involvement with the potential audience, and by their empathy with the main character of the story. The written narrative at this level is more compact, less redundant, and flows more smoothly for its lack of repetition, repair and excessive use of certain conjunctions. At the clause and sentence level, the difference resides not so much in the complexity of syntax as in how subjects encode and package information and events into longer utterances. While speakers keep their narration short and simple by coding events mainly into single-proposition clauses, writers select and integrate events into multi-propositional utterances and keep redundancy and wordiness to a minimum. Moreover at the word-level the oral narrative is frequented by hedges or hedge-like expressions, and characterized by a colloquial and colorful component of vocabulary, contractions, and clippings; the written one is better planned and executed with more precise diction and little colloquialism. In addition, similarities are found in discourse organization and episode building. The present study argues that while the similarities are attributed to the storytelling genre and narrative nature of the discourse, the differences arise mainly from distinct mental processes and mechanisms underlying speaking and writing.
1. Introduction For the past few decades, there has been a considerable body of research in various fields, such as linguistics, psychology, philosophy and education, dealing with the relationship between spoken and written language, especially similarities and differences between the two modalities. Although there is little agreement on the salient characteristics of the two modes, the general consensus is that they are different since they represent different ways of communicating and offer different ways of knowing and of reflecting on experience. They serve distinct functions and purposes in a discourse community, in time and space. They utilize different contexts. While oral language is typically associated with conversation that is produced and processed in the context of face-to-face exchange, written language is typically associated with the language of books and explanatory prose. Oral language is generally characterized as informal, interpersonal, and narrative-like with prosodic cues, deixis and paralinguistic devices readily available, and written language is considered formal, planned and expository-like, with limited reciprocity between the writer and the reader (Horowitz & Samuels 1987). Nevertheless, there is much variation and overlap in this simple dichotomy between oral and written discourse, depending upon the purposes for which they are used and the audience they serve.
In discussing that thought process of literate people may be different from those of oral people, Ong (1982) argues that there is little mental energy and capacity for innovation and exploration in oral cultures, but the text frees the mind of conservative tasks, that is, of its memory work, and thus enables the mind to turn itself to new speculation” (p. 41). Indeed, research on spontaneous speech reveals the relationship between the cognitive complexity and the actual production of speech output that conveys this set of thoughts. Greene and Capella (1986), for example, found that cognitive complexity increased at the juncture of ideational boundaries and is associated with a decrease in speech fluency. As a result, pausing and hesitation phenomena are often frequent in demanding tasks like evaluations as opposed to simple describing at choice point. The pressure and stress of speaking is more acutely felt by second language learners, who often consider writing an easier and less nerve-racking task. MacIntyre and Gardner (1989), for example, asked foreign students to write brief essays about language learning experience that they had found either positive and relaxing, or stressful and anxiety-producing. The results show that almost all of the anxiety-producing experience had had to do with trying to speak. In general, speakers are cognitively more constrained than writers because of the limited capacity of mental resources available for speaking, whereas writers are normally not pressed for on-line production.
Furthermore, cognitive psychology has offered an abundance of research that addresses the processing of speech and writing, and the short-term consequences of exposure to oral versus written discourse (see Jahandarie 1999 for details). There are several important findings on the comprehension of oral and written language at various levels of discourse. First, for sentence comprehension, reading seems to have a few advantages over listening at the stage of sensory coding and also through the activation of the faster orthographic code at the stage of lexical access, while listening has the advantage of direct access to the echoic memory with its much longer duration which is invaluable in the comprehension of connected discourse. Second, the relative consensus in the research on the retention of auditory versus visual stimuli seems to maintain that the two stimuli are retained in different memory stores and employ different memory mechanisms. Further, different modality superiority in recall performance has been found between people with different scripts in their written language. Third, quite a few studies have made direct comparisons between listening and reading of connected discourse, most of which consist of participants of elementary school children, who have not yet become skilled readers. The general pattern of findings among younger and poorer readers indicates a comprehension and recall superiority for both listening and oral reading over silent reading and reading-while-listening, while with more skilled groups of readers, silent reading becomes superior to listening and oral reading.
The large body of evidence from various fields of research has made it clear that major differences exist between speaking and writing, as demonstrated by many dissociations in the way the mind deals with the two language modalities. The different cognitive processes underlying spoken and written language would inevitably manifest themselves in different textures represented by the two types of discourse, as if they were the product of a different weave, with fibers of a different yarn (Halliday 1987:60).
2. Aim of the present study It is clear that the majority of evidence from separate lines of research indicates the existence of the differences between spoken and written discourse. Nevertheless, the comparison between the two modalities is usually either done between typical speech (e.g., dialogue or conversation) and typical writing (e.g., explanatory or expository prose), or between two different genres (e.g., narratives and academic papers). While distinct patterns have emerged from these comparisons, the characterizations may partly be due to the genre differences because patterns of discourse such as rhetoric structures, attribution, adversative, covariance, and response, etc. do not work in the same way across readers of various age groups and grades and across text topics (Horowitz & Samuels 1987). It would be more sensible and beneficial to examine oral and written language of the same genre and style to discover differences and similarities between the two modalities, where other variables of discourse are kept constant.
The present study aims to explore contrasts between the two modalities of the same type of discourse—a narrative account of a video-clip, which was produced under the same controlled conditions by subjects of homogenous backgrounds. This limited and controlled comparison would help us better understand differences and similarities between oral and written language in a stricter sense because other apparent differences such as audience, purpose, content, discourse context, discourse structure, etc. are kept minimal in our oral and written narratives.
There are two central issues to be presented and explicated in the study: We argue that, first, distinct mental processes found between speaking and writing would inevitably manifested themselves in the production of spoken and written narratives, and second, similarities and differences between the two modalities would not only be present within a particular language but also be found across languages because of the universal characteristics of narrative discourse and human cognition. On the one hand, oral and written narratives are posited closely together on a speaking writing continuum with regard to structure, level of formality, message content, and genre of the text, and on the other hand, they involve distinct mental processes and mechanisms, regardless of their specific language. In this paper, we will examine and explore patterns of similarities and differences between the two types of narratives across two different languages, using narrative data elicited from native speakers of American- English and Mandarin Chinese.
The stimulus material consists of a 4-minute video-clip entitled The New Doorbell. The clip is a cartoon about a man who installs a new doorbell in his apartment and then waits anxiously for others to ring it. The cartoon is a silent color movie with background music; no written language ever appears on the screen except the title, which is shown at the beginning of the video-clip in both Chinese and English. This video-clip rather than some reading or listening material was selected because of the argument that English and Chinese speakers may process the scripts and oral speech in their respective language in different ways, and we wanted to make sure that the subjects of both languages composed the narrative from the same visual stimulus material. More importantly, producing stories in oral or written form from the same non-linguistic material would give us sufficient control over many variables that may influence the resulting narrative (e.g., the translation of an oral or written passage from one language to another), and at the same time afford the subjects flexibility in organizing and constructing narratives linguistically.
There are eight episodes in the story, defined and identified independently from linguistic material. Each episode involves some kind of change or changes (e.g., location, event, character) in the scene from the previous episode (c.f., Pu 1995). The episodes (EP) are summarized as follows:
EP1: A man walks into his apartment and installs a doorbell. He rings the doorbell outside the door and enjoys the music it makes.
EP2: The man hears footsteps, closes the door, hoping someone would ring his doorbell.
EP3: A young man comes by, stops at his door to tie his shoes and then leaves.
EP4: The man is disappointed, but he decides to sit down, reading a newspaper and waiting.
EP5: A little girl comes by, bounces a ball in the hallway and then leaves. The man is again disappointed.
EP6: A fat man comes to the apartment, knocks on the door, hears no answer, and leaves. The man is upset.
EP7: The postman comes and knocks on the door. The man gets angry and pounds on the door from inside. He finally opens the door, still trying to coax the postman into ringing the bell.
EP8: As the man closes the door, the postman leaves. The man is bitterly disappointed.
As in any other typical story, The New Doorbell consists of a beginning which introduces the main character and the setting (EP1) and the purpose of the story (EP2), a middle which unfolds the storyline (EP3-7), highlighted by a climax (EP7), and an end which brings the outcome of the story (EP8). The overall structure of the story is schematized below:
The New Doorbell
Beginning Middle End
Setting and Purpose Development and Climax Outcome
(EP 1 & 2) (EP 3 to 7) (EP 8)
The schema shows that the story is both linearly and hierarchically structured. It would be interesting to see how speakers organize the story information, what episodes they consider important, what less important and what marginal; how they linearly produce the hierarchical structure of the story, and what linguistic devices they use to achieve the story coherence. Furthermore, we were particularly interested in differences and similarities between the subjects performance in either oral or written task across the two languages.
The narratives were produced either individually (oral narratives) or in groups (written narratives). Before the video presentation, subjects were given a written instruction in their native language, which tells them to first watch the video-clip, and then describe what happened in the clip. Half of the subjects were instructed to describe the clip in oral form so that someone who hasnt watched it will be able to know what has happened in the clip by listening to your description. The other half read the same instruction except that they would describe the clip in written form so that someone who hasnt watched it will be able to know what has happened in the clip by reading your description. The written instruction does not describe the nature of the video-clip and the purpose of the study, nor did it mention the word story or story-telling. We hoped that the subjects were able to perceive, construct and describe the video-clip with as little preconceived notion as possible. The oral narratives were tape-recorded and later transcribed; the written ones were collected immediately after the subjects finished writing the story.
Seventy subjects participated in the narrative study. Forty were native American-English speakers from the University of Maine at Farmington, and thirty were native Mandarin Chinese speakers from the Central China University of Finance and Economics. The subjects were randomly assigned to two groups in each language: 20 in English Oral (EO), 20 in English Written (EW), 15 in Chinese Oral (CO) and 15 in Chinese written (CW). All subjects were undergraduates and about two-thirds were women.
3. Contrasting speaking and writing at various levels of discourse In general, the speakers of both groups across the languages produced the narratives in comparable ways in terms of episode selection and description, coherence-building, event sequencing, reference tracking, and inference making. Although the overall structure of the narrative across the four groups is quite similar, striking differences are found between the two modalities. The remaining sections of the paper will focus on how the subjects of both languages organized and constructed narratives, and what patterns of differences and similarities emerge between the oral and written narratives.
3.1 Overall organization
The speakers and writers of the both languages told the video story in surprisingly consistent ways. They did not merely represent the visual data, such as the movements of the objects or persons, but rather, an interpretation of the events (Loftus 1979). In other words, they constructed a meaning. While describing what happened in the video-clip, they were more concerned about the cause and outcome of the actions, and the purpose and explanation of the characters who performed the actions. They stayed on the main event line, focused on the actions of the central character, and paid less attention to information that are peripheral to the development of the storyline. Although not all episodes were described by every subject, all narratives exhibit the global structure of story construction: a beginning consisting of background information such as the setting and the theme of the story (EP1-2), a middle showing the development of the story with a climax (EP3-7), and an ending describing the outcome of the events (EP8). This global organization of the narrative discourse demonstrates that subjects, or rather, people in general, have expectations about how to tell a story. They take in visual stimuli, construct a mental representation of what they perceive, and encode it into a linguistically structured message. Even though subjects were not instructed to tell a story, they uniformly organized the visual data into a hierarchical structure with interrelated events, describing who, when, where, what happened and why in their narrative, be it oral or written. Hence in terms of structure organization and discourse coherence, oral and written narratives are strikingly similar because of the constraint of story frame, i.e., the speaker/writers expectations about how stories should be told and the fulfillment of the expectations (c.f. Tannen 1993).
The constraints of story frame are further evidenced in the selection and recall of episodes in all narratives across modalities and languages. Prior research has demonstrated the existence of episodes as chunks in narrative memory (Guindon and Kintsch 1982, Pu 1995): comprehenders easily perceive the episode boundaries after reading a written story, listening to a spoken story, reviewing a nonverbal picture story, or watching a movie without dialogue. Our narrative data lend support to the mental representation of story as episodes, which is revealed by the subjects organization of the story. In general, episodes are formulated explicitly and marked. Both speakers and writers used adverbs, adverbial phrases or noun phrases to mark the beginning of an episode such as next, then, after that, at this point, finally, eventually, the first person, the second person, the last person, etc., and episodes are often separated by indentation and/or a blank line in written narratives. Moreover, episodes as memory chunks are shown by the fact that speakers sometimes displaced an episode out of its sequential order, and some overtly mentioned, afterwards, a missing episode that somehow slipped their memory.
Furthermore, it is of interest to see how subjects maintained the main storyline by selecting major episodes or important events in an episode because a narrator cannot recount every detail. In general, subjects had an implicit agreement with what are important elements of the story and developed the story with these elements. They were very consistent in their selection, omission, and description of the episodes. Subjects across the four groups described EP1 and EP7 in detail because these are the backbone of the story, delineating the setting and the climax. On the other hand, the majority of the middle episodes EP3-6 are rather sketchy because they are not the highlight of the story, which may be touched upon but not worth detailing. Sometimes an episode of this section, most often EP4 is omitted, which describes the man reading a newspaper in his apartment while waiting for someone to ring the doorbell. Obviously, EP4 is not crucial to the development of the storyline because it is not directly related to the purpose and the outcome of the events, and hence its omission does not sacrifice the coherence and completeness of the story. What is more interesting is that when EP4 was actually being described, more than half of the subjects in each group displaced it in their recall of the story. This seems to reflect how speakers construct a mental representation of the story, which consists of the skeleton of the story, only to be fleshed during verbalization with macropropositions and micropropositions at various levels to make the story structured and coherent. Since EP4 figures less prominently (less consequential) in the development of the story (i.e., in the mid-section) and its precise order may not be remembered as accurately as a semantically or thematically more important episode such as EP7. Therefore, subjects verbalized EP4 in various ways: unloading it at the beginning of the middle section, holding it until before the climax, or omitting it entirely. Compared with EP4, other important episodes such as EP1 (the setting), EP7 (the climax), and EP8 (the ending) were always recalled and in their accurate order because of their prominence in the story and hence memory.
The episodic organization of the story again shows how subjects achieve discourse coherence by organizing information and events into episodes. These episodes are related hierarchically to reveal the purpose of the story and linearly by the temporal order of events to achieve discourse continuity. During the process, important information and events are always included and less vital elements are more prone to be omitted or confounded.
We have demonstrated that at the global level of discourse organization, subjects performed quite uniformly across modalities and languages in telling a story. Nevertheless, striking differences are found at all levels of narrative discourse between the two modalities. It is to the patterns of differences we now turn.
3.2 Patterns of differences
3.2.1 Features of words and phrases
The first major difference emerging in the English data is that of verb tense between the two modalities: While 12 subjects in oral narratives used past tense to tell the story, 15 in written form employed the present tense, and the use of present and past tense is found to be associated with the modality (t2 = 6.46*, p0.01). The difference in verb tense may be due to the nature of the oral versus written story-telling. In EO condition, 13 out of 20 subjects started their narrative with a comment that they just watched a video/movie clip, and this very first sentence has set up a time frame for their subsequent story-telling, i.e., what happened in the video. In EW, however, very few of the subjects mentioned the fact that they were about to tell a story of a video clip they just watched; sixteen out of twenty subjects started the narrative directly with the central character of the story, e.g., A man ... or There is a man ..., and then continued almost always with the present tense description throughout the written narrative. The typical beginnings of the oral versus written English narratives are presented in (1)-(4) below.
1) EO: S9
In the video we just watched, we saw a man ...
2) EO: S10
We watched a video ... called the new doorbell. It was about a man ...
3) EW: S5
There is a man who lives in ...
4) EW: S7
A man walks up two flights of stairs, and then ...
While speakers were recounting past events, writers seemed to be describing what was happening at the moment. Moreover, one subject in EO and one in EW switched the tense from past to present in the middle of their narrative, both of whom began their story with the mention of the movie clip they just watched, followed naturally by past tense description, only to shift to the present tense at EP7. The shift is illustrated in (5) and (6) below.
5) EO: S18
EP1: Umm, I just watched a video .. ah about a man ...
EP7: Finally it ends by this ... it appears to be a policeman type of character, couldve been a mailman, and I dont know, ...
6) EW: S19
EP1: The film clip started with a man ...
EP7: Next you hear a delivering truck outside. The delivery man comes upstairs ...
Interestingly, both subjects switched the tense in EP7, where the story takes a dramatic turn. It appears that the subjects subconsciously pulled the audience back to the present at the climax of the story to present a more vivid imagery of what was happening here and now, rather than what happened there and then. The use of present tense verbs as a variant of the sequential past is discussed in Givón (1993), who argues that this usage involves a manipulation of the pragmatic perspective of the discourse, which makes the events somehow more vivid or immediate, “as if the narrator invites the hearer to be present on the scene, observe the action from closer quarters, be more emotionally involved (p. 166). It seems that by using the present tense to tell the story, writers were trying to compensate for their detachment from readers and actively engage them in their reading process.
Although Chinese does not have overt markers for verb tense, subjects did begin their narratives in the two modalities differentially to a certain degree. Eight out of fifteen subjects in the oral task started their narrative with the mention of the video/cartoon, suggesting that they were telling the story of a video which they watched previously. On the other hand, fourteen subjects in the written task opened their narratives, as did their English counterparts in EW, with some form of introduction of the central character, and then continued to report the events in their temporal sequence. Consider (7), an oral opening, and (8), a written one.
7) CO: S9
嗯，这个动画片的 名字 叫“新装的 门铃”…
um, this cartoon of name call new-install doorbell.
Umm, the title of the cartoon is The New Doorbell. ...
8) CW: S2
一个男人住 在一幢公寓 里, …
a man live at a apartment in
A man lives in an apartment, ...
Another difference found at the word/phrase level between the oral and written narratives is the occurrence of hedges and hedge-like expressions such as kind of/kinda, sort of, anyway, etc, which are used almost exclusively by the oral group of both languages. For example,
9) EO: S3
The delivery man kinda looked at the doorbell, and then walked away.
10) EO: S7
Then a short little man came and knocked on the door, and this made him kinda angry.
11) CO: S2
这 好象是一个中年 人，是吧?
this seem-like is a middle-age man, isnt it
This seems like a middle-aged man, doesnt it?
12) CO: S12
他可能 有点儿 失望吧，
he maybe somewhat disappointed
He is kind of disappointed.
Such expressions suggest that although the speakers settled quickly on words such as angry, frustrated, disappointed, middle-aged, etc., they seemed uncertain of their lexical choice. Chafe and Danielewicz (1987) attribute the hedge use to the limited lexical repertoire available to the speaker, who is not completely satisfied with his/her lexical choice, yet has no time to ponder on a better word because speaking is done on the fly. They argue that speaking calls for more expenditures of cognitive effort, and hence speakers tend to operate with a narrower range of lexical choices than writers. As a result, the vocabulary of spoken language is more limited in variety, regardless of the kind of speaking involved. While the argument is intuitively appealing, it is not evidenced in our study. Although writers of both languages used few hedges such as those shown in the above samples, their choice of vocabulary is very much the same as those of the speakers even if they have the leisure to dip into the rich storehouse of literary vocabulary, search for items that will capture nuance (Chafe and Danielewicz, 1987:88). For example, when describing the mans disappointment, subjects across the four groups used the same set of adjectives such as upset, angry, frustrated, disappointed, disheartened, sad, and mad, accompanied sometimes by intensifiers or quantifiers such as very, extremely, a little, more, and so. It seems, at least, for the genre of story-telling, the range and level of vocabulary are quite similar between the two modalities, which is in general simple and of high frequency.
Why, then, do speakers seem less certain on their lexical choice than writers, given the fact that they practically use the same set of words. We argue that what speakers hesitated about but had no time for is not so much the availability of vocabulary as the verification of what they had actually seen prior to the story-telling. For example, the speaker in (9) was making a simple statement that the postman just looks at the doorbell button before he leaves. The subject had no trouble with the choice or preciseness of the verb look (at), but she wasnt sure, at that moment of the on-going production process, whether the postman actually looked at the doorbell or not. She couldnt afford to think more about it, but used a hedge instead to show her uncertainty. Similarly, (10), (11) and (12) above reveal that the speakers knew that the man was angry, frustrated or disappointed but they were not sure of the degree of his anger, frustration or disappointment. In order to maintain the quick and smooth flow of production, they had no choice but to use a hedge to mark their state of mind on that particular description. Of course, speakers can pause, or false-start to revise what has been said or explicitly mention their uncertainty, but too much fumbling is harmful to effective communication on the one hand, and acknowledging Im not sure/I dont know is damaging to their credibility as a storyteller on the other.
In written narratives, however, the use of hedges and false start such as those shown in (10)-(13) above almost never occurred because in writing, with or without editing, one always has more time for language processing. Writers usually plan a clause/sentence ahead before they actually write it down, and they have time to resolve uncertainly and avoid hesitation before producing a word or phrase. For example, while many speakers reveal their hesitation about the degree and intensity of the mans anger, frustration or disappointment by using the hedge kind of/kinda in English and maybe, somehow in Chinese, writers had more time to think about, decide on, and hence better describe the man’s emotion without explicit hesitation. Consider the following.
13) EW: S3
The man inside is getting angry.
14) EW: S4
Only this time the man grows very angry, and bangs on his door from the inside.
15) CW: S7
王先生 打开 门，那人 早已 走了，王 先生 十分恼火
Wang Mr. open door that-guy early go Wang Mr. very frustrated
When Mr. Wang opens the door, that guy has long gone. Mr. Wang is very frustrated.
The postman walks off, leaving behind the extremely disappointed owner.
It seems that writers are aware, consciously or not, of the permanency and the formality of the writing and try to avoid hesitation and uncertainty in their narratives. What is written down exists (more permanently), spread out on the page. Informal words such as kind of/kinda dont usually belong to the written form unless one wants to mimic the oral language on purpose. In addition to hedges and hesitations, informal wordings, vague expressions, partial words, clippings, and contractions occur in English oral narratives much more than in written ones. The following examples illustrate the informal style of oral storytelling.
17) EO: S1
Then he sat down and was kind of bummed out.
18) EO: S3
And he finally opened the door, and the delivery man got him to sign the thing,
19) EO: S15
He rings it and he runs inside and starts dancing around the room cause hes excited.
20) EO: S5
So the man was standing in the room thinking: Well, isnthe gonna ring the doorbell?
Spoken narratives commonly employ contractions, which is a feature much like the use of informal, colloquial vocabulary. There are altogether 106 cases of contractions in the spoken English narratives whereas only 20 are found in the written ones (the difference is statistically significant; t=3.73*, p0.05).
The Chinese oral narrative has similar colloquial components, which is shown in the following examples.
21) CO: S1
哎哟, 底下 摩托 声 响了。哎哟，这 肯定是送信的 来了
Excl down motor sound rise Excl. this must be send-letter-man come
Ah, a motorcycle is heard downstairs. Ah, this must be the mailman coming.
22) CO: S11
咔咔咔， 听见脚步声 来了，当时 就立马 兴奋起来
onomatopoeia hear foot-steps come at-once just right-away excited up
Ka-ka-ka, (he) hears footsteps coming and gets excited right away.
In (21), the subject used送信的(‘letter-carrier’), a colloquial term for 邮递员 (postman), and the subject in (22) preferred 立马 (right away) over a more literary term of 迅速 or立刻, all meaning immediately or instantly, which frequently occur in the written narrative. However, colloquial words sometimes carry more force than their literary equivalents in expressing the meaning orally, assigning a premium to freshness (Chafe and Danielewicz 1987).
3.2.2 Characteristics of longer utterances
At the clause or sentence level, oral and written narratives are both comprised primarily of main clauses, usually simple and short. The similarity in clause length and syntax reflects the nature of narratives, be they oral or written, which are modeled on the storytelling genre and whose effect depends on interpersonal involvement between the speaker/writer or the character and the reader (Tannen 1984).
On the other hand, complex sentences are found in both modalities, of which adverbial and relative clauses occur more frequently than other dependent clauses. The following table shows the distributions of these two types of subordination in each of the four tasks.
In general, there is little difference in frequency of distribution among the four tasks except that of relative clauses between Chinese oral and written narratives. However, there are important differences between the two modalities with regard to the structure and function of complex sentences beyond what the numbers suggest.
Let us first look at English adverbial clauses, which consist mostly of clauses of time and of cause/reason. There are two alternative structures of English complex sentences with regard to the relative position of main clause (MC) and its subordinate clause (SC), where MC+SC being the unmarked structure, and SC+MC the marked one (Beaver 1970, Thompson 1985). We find that there is a difference between speaker’ and writers’ choice of a particular structure when they relate events with temporal clauses. In the oral task, speakers encode events into a complex sentence in the order they occur in the story, i.e., event 1(E1) + event 2(E2), regardless whether it is MC+SC or SC+MC. For example,
23) EO: S2
He showed the postman the doorbell before he closed the door, hoping the postman would ring it.
24) EO: S12
When he pressed it, it played music.
25) EO: S5
He danced to the music until he heard footsteps coming upstairs.
In each of the three utterances, the order of the two clauses mirrors the temporal sequence of the events, which is typical of narrative discourse and an important way of maintaining discourse coherence and the storytelling frame (Tannen 1993). Indeed, of 27 complex sentences with temporal clauses in EO, 23 (85%) code the sequence of E1+E2, of which 12 are in the order of MC+SC and 11 of SC+MC. It seems that iconicity of discourse overrides markedness of syntax in the oral production. Writers, on the other hand, appeared to manipulate the alternative position of MC and SC to serve different discourse functions. There are altogether 40 temporal clauses in the written data, 67% of which (27 out of 40) are in the order of MC+SC. While most MC+SC utterances reflect the temporal order of events within an episode, all 13 SC+MC utterances are found at the beginning of an episode, marking the boundary and signal the advent of the new episode. For example,
26) EW: S6
EP7: After the man waited some more, he heard another person come up the stairs.
27) EW: S10
EP3: When he hears someone walking up the stairs, he gets excited.
28) EW: S11
EP2: Once he is satisfied with the new sound, he goes back into his apartment.
Nevertheless, the marked structure of SC+MC in the oral data is not found to be associated with the function of separating episodes. The use of marked structure to serve specific discourse functions has been observed and discussed in several prior studies (Carpenter & Just 1975, Gernsbacher 1990, Thompson 1985, Pu & Prideaux 1994), and our experimental study further suggests that oral and written discourse may impose different effects on the use and function of marked structure.
The difference in the encoding of temporal clauses between the two modalities is again a manifestation of cognitive constraints underlying speaking and writing. On the one hand, it makes perfect processing sense that speakers should rely more heavily on the temporal sequence of events than writers during the storytelling because they are more constrained in cognitive resources during narrative production; a chain of events that are narrated in a coherent, timely sequence is easier to encode, store in memory and retrieve since such an iconicity is cognitively less demanding (Gernsbacher 1990, Givón 1993). Writing, on the hand, provides more time and ultimately more cognitively resources. Although writers also frequently code events in their sequential order of E1+E2 to maintain discourse coherence (which is, anyway, typical of narrative discourse), they can nonetheless afford to manipulate syntax, e.g., to use a marked structure in this case, to indicate thematic discontinuity.
Another point of interest concerning adverbial clauses is the use of cause- or because-clause between oral and written tasks. Whereas the occurrence of such clauses in oral task outnumbers that in written task (21 vs.11), they exhibit different degree of cause-effect relationship between the two modalities. In the written narrative, 10 out of 11 clauses are used exclusively to describe one scene in EP6, expressing the reason for the mans refusal to open the door (7 in total), or the cause of his anger and frustration (3 in total). Consider:
29) EW: S8
EP6: The person knocked but the man didnt answer because the other person didnt ring the bell.
30) EW: S12
EP6: The man gets very frustrated and refuses to open the door because he wants him to use the doorbell.
31) EW: S18
EP6: This angers the man inside because he wants him to use the new doorbell.
The writers seemed much obliged to explain the mans refusal to open the door when someone knocked, an unusual event that doesnt quite meet one’s expectation and hence prompts the question of why he didnt open the door. The connection between the reason and the affected event, i.e., the mans reaction to the knock on his door, was deemed necessary and therefore a cause/reason clause was used. In the oral narrative, on the contrary, only 6 out of 21 because-clauses describe such a necessary causal relationship between the two clauses, others are used in a much looser sense in expressing reason and cause. Many of these clauses are found to describe scenes as shown in (32)-(35) below.
32) EO: S4
He didnt ring the doorbell cause I dont think he saw it.
33) EO: S13
And the guy inside was getting very angry because he started pounding on the door.
34) EO: S15
The man stands up and puts his ear by the door cause hes waiting.
35) EO: S18
He was very anxious, yknow, cause he could hear the person outside.
In the above utterances, the invocation of reason is subtle and even opaque. For example, in (32), the reason for the affected event, i.e., ‘he didn’t ring the doorbell’, is subjective. In (34) and (35), however, the reason clause (‘he’s waiting/he could hear the person outside’) does not explain directly and explicitly why ‘he puts his ear by the door/he was very anxious’, which is left for the listener to infer. In (33), on the other hand, the reason and effect seem to be encoded reversely in the main and adverbial clause, i.e., the clause he started pounding on the door is not a reason for why ‘the guy was getting angry’, but vice versa, ‘the guy inside started pounding on the door because he was getting very angry.’
It is not surprising that speakers seem to be more liberal in their account for the reason-effect relationship than writers because speaking is done on the fly and speakers have little time pondering on the effectiveness and appropriateness when conjoining clauses. Anything that strikes the speaker as relevant may be cited as reason (Givón 1993, p. 300). Writing, on the other hand, is slower and editable, which provides writers with ample opportunity to construe sentences to express the relationship such as cause-effect more precisely.
Likewise, Chinese speakers and writers used adverbial clauses differentially in their narrative production. Although adverbial clauses occur in oral narratives as frequently as they occur in the written ones, they differ in type and variety between the two modalities. Let us first examine temporal clauses. In the oral task, most of these clauses (53 out of 59; 90%) take, singularly, the linking word ‘after’ at the end of the clause. In the written task, however, though the temporal clause still rank the highest in frequency, half of them (22 out 44) indicate time with a variety of linking words such as 一… 就 (‘as soon as’), 当… 时候 (‘when’), 一边 … 一边 (‘while’), etc. Consider the following oral versus written utterances.
36) CO: S3
回到家 以后, 他 把 外套 脱了,
come home after he OM jacket take-off
After coming home, he takes off his jacket.
37) CO: S4
他 把 门 关了以后， 邮递员 转身 走了。
he om door open after postman turn-body go
After he closes the door, the postman turns and leaves.
38) CW: S14
他一 回到 家 就开始 装 门铃。
he as soon as return home just start install doorbell
As soon as he returns home, (he) starts to install the doorbell.
39) CW: S15
他一边 看 报纸， 一边 竖着耳朵听 外面的动静。
he while read newspaper while prick ear listen outside sound
While he reads the newspaper, he pricks his ears to listen to any sound outside.
Moreover, while adverbial clauses are used almost exclusively to encode time in oral data (59 out of 60), about 30% in the written task are found to relate to their main clause thematically in other ways, indicating cause, reason, purpose, and concession. For example,
40) CW: S5
虽然 他等了很 长 时间, 但 他并不气馁。
though he wait very long time but he not discouraged
Though he has waited for a long time, he is not discouraged.
41) CW: S6
因为 门铃 装得 太 高了，他够 不着。
because doorbell install too high he reach not
Because the doorbell is too high, he can’t reach it.
The differences between speakers and writers in their choice of adverbial clauses with regard to type and content again reflect cognitive activities underlying discourse processing. On the one hand, since Chinese temporal clauses most commonly precedes their main clause, the structure of after-clause+MC clearly encodes two events in the sequence they occur in the story. Consequently, temporal clauses in CO are mostly of after-type because the Chinese speakers, like their English counterparts, prefer iconicity in discourse processing which is cognitively less demanding. On the other hand, adverbial clauses other than those indicating time are almost non-existent in CO because construction of utterances more complex in content and theme must call for the expenditures of some cognitive effort, which speakers constantly lack but writers have. In contrast, the written narrative witnesses adverbial clauses of a greater variety.
Another striking difference between the two narrative tasks in Chinese is the distribution of relative clauses: only 10 in CO but 58 in CW. The rarity of relative clause in the oral narrative may be attributed to how cognitive strategies interact with the structural characteristics of the clause. Generally speaking, Chinese relative clause precedes the head noun, and moreover, it sometimes separates the head noun from its classifier or demonstrative such as ‘a’, ‘some’, ‘this’, or ‘that’, etc. These conditions frequently render Chinese relative clauses to be an embedded clause in the sentence. For example,
42) CO: S7
然后心里 就 只 想着 那个[他 刚 装上的] 门铃
then heart-in still only think-of that he just install door-bell
All he could think of then was the doorbell that (he) just installed.
43) CW: S1
一位 [戴着墨镜的] 男 青年 走到 他 门前 停了下来
a wear dark-glass male youth walk-to his door-front stop down
A young man who wears dark glasses stops at his doorstep.
Our data show that of all 68 Chinese relative clauses, most (94%; 64 out of 68) are center-embedded. The particular structure of Chinese relative clause poses difficulty for its processing because of our limited cognitive capacity. On the one hand, embedded clause disrupts main clause, and on the other hand, relative clause, a relatively long modifier, preceding its head denies a quick access to the head. Both conditions put greater demands on short-term memory since information about the incomplete main clause must be retained in memory while the embedded clause is being processed (Slobin 1973, Prideaux and Baker 1986, Just & Carpenter 1992, inter alia). Hence in the quick flow of oral production, Chinese speakers try to avoid using relative clause construction as much as possible in order not to overburden themselves. The cognitive constraints underlying spoken language affect writing to a much less extent and hence writers can afford to use center-embedded relative clauses more frequently.
Nevertheless, the pattern is reversed in English narratives, where speakers produced more relative clauses than writers did (48 vs. 34). This is so because English speakers may simply avoid producing embedded relative clause by putting it at the end of sentence. For example,
44) EO: S3
Then he pressed it again, he went inside and danced to the song [it played].
45) EO: S8
But the man comes back and grabs the postman [who is about to leave].
Both of the utterances contain a relative clause that modifies the grammatical object of the main clause and that is at the sentence-final position. Relative clauses such as (44) and (45) do not interrupt their main clause, which can be discarded from short-term memory before the relative clause is processed, and are thus less costly in terms of mental resources. Indeed, of all relative clauses produced by English speakers, the majority (41 out of 48) are sentence-final ones.
Another important difference between oral and written narratives at the clause/sentence level lies in how speakers and writers package and encode events in sentences. While constructing narrative orally, subjects tend to put sequences of events in consecutive yet separate clauses or sentences, usually one proposition per clause. In written narratives, on the other hand, clauses and sentences are more compact, consisting multi-propositions per clause. The contrast is illustrated in the following English and Chinese examples, where each pair, (a) and (b), describes the same event.
a) EO: S11
Next a little girl comes up, and shes bouncing a ball,... And she bounces the ball outside his door, and then leaves the scene. She doesnt ring the doorbell.
b) EW: S14
A little girl bouncing a ball approached the door and walked away without ringing the doorbell.
a) CO: S15
刚开始的 时候有 一个人， 他 好象是 回 家吧。然后可能是 天 黑了，他
at beginning time have a person he seem-like go home then maybe day dark he
上了二楼 ， 一进门…… 推 开 一个门，开了 灯 之后，他 就
go second-floor upon-entering-door push open a door turn-on light after he just
安了一个门铃。 先 在里面 装了一个盒子之类的东西，然后跑到外面去，把
install a doorbell first in inside install a box like thing then run outside om
外面的一个摁的 东西，一个按钮 也 安好了。然后他就自己感觉挺 不错的。
outside a press thing a button also install then he self feel very fine
然后 他就 按了一下 门铃，然后那个门铃 发出了音乐声，他自己从 门
then he just press once bell then that doorbell emit music he self from door
外 转 到门 里面，觉得挺 高兴的。
outside waltz to door inside feel very happy
At the beginning, there is a man, he seems like coming home. Then it’s getting dark. He goes upstairs to the second floor, upon entering the door, … pushes open a door. After turning on the light, he starts to install a doorbell. (He) first installs a box-like thing inside (the door), and then goes outside to install a presser, a button. Then he feels satisfied. Then he rings the doorbell. Then that doorbell plays music, he himself waltzes from outside to the room, feeling very happy.
b) CW: S8
一天，一个带 眼镜的 中年 人 回到 家，兴高采烈地在门上装了 个
a day a wear glasses middle-age man return home excitedly at door install a
门铃。 装 好 后 他一按 按钮，立即 悦耳的音乐 便 响 起来
doorbell install well after he press button immediately pleasant music then soud up
代替了以往 单调的敲门 声。 他 完全 沉浸 在新装门铃的喜悦之中。
replace usual dull knock sound he thoroughly immerse in new doorbell happiness
One day, a middle-aged man who wears glasses comes home, and excitedly installs a doorbell on the door. Upon completion he presses the button, and immediately, delightful music plays instead of the usual, dull knocking sound. He is thoroughly immersed in the joyfulness that the new doorbell brings.
(46a) and (b) above describe the same sequence of events in EP5 about the little girl who passes by the mans door without ringing his doorbell. The subject who told the story orally coded the episode almost scene by scene in 5 separate clauses, each of which consists of only one proposition, whereas the subject who wrote the story encoded the episode in two clauses, each of which is comprised of more than one proposition. Likewise, the subject in (47a) employed utterances of single proposition to describe the opening scene of the story, but the writer in (47b) encoded the same episode in utterances of multi-propositions with fewer words. Moreover, the writer used a number of adverbial phrases of time and manner such as ‘excitedly’, ‘immediately,’, ‘thoroughly’, etc. to describe simultaneity of events and/or the character’s feelings and state of mind without being wordy or redundant. The contrast between oral and written descriptions is stark in this respect: speakers in general prefer single-propositional utterances since those utterances are easier to store and retrieve in language processing than multi-propositional sentences, as has been found by several prior experimental studies (Kintsch and Keenan 1973, Kintsch and Glass 1974); written passages are more compact and more complex in terms of the number of propositions per clause and the strategic deployment of present and past participles (in English) and adverbial phrases (in Chinese) .
3.2.3 Structure of discourse
The difference in discourse planning between speaking and writing is further evidenced beyond the clause/sentence level, and the overall narrative is in general full of signs and cues which give away whether it is spoken or written.
The first difference between the two modalities in English is the way speakers and writers conjoin sentences and clauses. Spoken narrative consists mainly of simple sentences joined together with the coordinating conjunction and, which serves not only to link between clauses and sentences but also to sequence events in their temporal order. Speakers try to avoid complex interclausal relations because elaborate syntax evidently requires more processing effort than speakers can ordinarily devoted to it (Chafe and Danielewicz 1987: 103). There are altogether 431 and used by the English speakers, more than 20 per narrative, which are located at the beginning of a clause. The following passage is taken from the English oral data, where the alphabets (e.g., a, b, c, etc.) are inserted for the convenience of discussion.
48) EO: S6
a) A man comes home from work or from the store, and he has just bought a doorbell. b) He takes off his coat and installs the bell inside his apartment, and um, then goes to the outside and installs .. the ringer. c) Um, he goes .. and pushes the button, and listens to it ring, and gets all excited. And he pushes it again as hes going inside. And hes dancing around, and hes all happy about his new doorbell. d) And then he closes the door.
In the above passage, almost every clause or sentence starts with the conjunction and, which the speaker used to chain together a sequence of events or actions, expressed by simple sentences. What is more interesting, however, is when such and is not used. Our data shows that whenever a chain of events ends or there is a weak link in the chain, speakers would either prefer not to use and, or rather, use and then to emphasize a transition to a new sequence of events. In (48), for example, sequence (a), with two clauses joined by and, introduces the background information that the man got home with a new doorbell. In sequence (b), the narrative shifts to the foreground information of the man installing the doorbell, and the shift is marked by the disappearance of the ubiquitous and, indicating a slight thematic discontinuity in the sequence of events. Within the sequence, however, all clauses are linked by and. Sequence (c) shifts again from describing the process of installing the doorbell to the man’s joyfulness upon its completion. At the juncture of the transition, and is not used but it recurs within the sequence. Sequence (d) starts with a new scene as the man gets back to the room, and the beginning of the sequence is marked by and then.
In contrast, the use of and in written English narratives is more conventional, i.e., it is used as a coordinate conjunction to link two or more clauses of parallel functions instead of as a chaining device sequencing up events of discourse. For example,
49) EW: S17
This movie clip was taken from The New Doorbell. A man in a white shirt and glasses walked up two flights of stairs to his room and opened the door. He turned on the light and used a screwdriver to install a red doorbell. He placed the doorbell very high on the door, reached up, and pressed it several times. He danced to the tune it played.
In (49) all three and are used to conjoin main clauses, and there is no and appearing at the beginning of a sentence. Comparing spoken with written data, we see other differences between the two modalities. For example, in addition to the excessive use of and, speakers often used fillers, repairs, repeats and pauses between phrases, clauses and sentences such as um, uh, okay, etc. to show that there is not sufficient time for discourse planning and they are very much constrained by their cognitive capacity. English written narratives, on the other hand, seem to be better planned and writers have time to think of what they want to put in the sentence and the relations between sentences before producing it. The narrative is free of fillers and repeats, although there are occasional repairs, shown by the scratch of a word or a phrase (found only in 3 narratives). In fact, the sentences of the written narrative are not necessarily longer nor do they contain more subordination than those of the oral ones, but their construction nevertheless shows more careful planning and succinctness.
Chinese narratives exhibit similar types of differences between the two modalities as do the English narratives. In oral narratives, clauses and sentences are linked most frequently by the conjunctions然后(then or and) and 结果/所以 (so): 187 in total, a little more than 12 per narrative. Like their English counterparts, Chinese speakers often repeated, repaired and employed fillers such as 好(well), 这个/那个(uh, ‘um’), etc. to buy time for the production of next utterance, whereas the written narrative flows more smoothly and shows more varied linking devices between clauses and sentences. The contrast is shown in the following examples:
50) CO: S5
下 一个人 .. 一会儿来了 个中年 人, 比较 胖, 挺 矮的。然后
next a person a while come a middle-age man a-little fat rather short then
他 在, .. 肯定又是 找 他。结果 .. 不过 他 没 按 门铃, 而是在那个
he at must again look-for him so however he not press door-bell but at that
门上敲了 一阵儿, 敲了一阵儿, 那个屋 主人呢, 他也 可能 很 失望
door knock a-while knock a-while that room owner he also maybe very disappointed 他，他，他就 没有什么，就没有这个回， 那个.. 开，去 开 门，还 希望
he he he just not what just not this return that open go open door still hope 那个外面的人 按 门铃
that outside person press door-bell
Next person .. after a while comes a middle-aged man, a little fat, rather short. Then he .. must be looking for him. So .. but he doesnt press the doorbell, yet knocks on that door for a while, for a while. That apartment owner (ne), he may be very disappointed. He, he, he doesnt do, uh, doesnt uh, return, uh, open, doesnt go open the door, hoping uh the man outside would ring the doorbell.
51) CW: S3
不一会， 一个矮矮胖胖的男士B直奔 A家。 也许 上楼 对他来说
not awhile a short fat man straight A home maybe go-upstairs to him say
不容易，气喘吁吁。 这 气喘声 也 没 能 逃过 A的耳朵。他
not easy out-of-breath this hard-breathing also not able escape As ears he
兴奋地 站在 门 后， 就 等 音乐 响起。可 不知是B没有 看见，还是
excitedly stand-at door behind just wait music sound but not be B not see or
因为 太 矮了够 不着，只 见 他 挥起拳头咚咚地 敲 门。
because too short reach not only see he raise fist onomatopoeic knock door
男士A对男士B 没有丧失信心，耐心地 等著 他 发现门铃。
man to man not lose faith patiently wait him find doorbell
After a while, a short, fat man B comes straight to As apartment. Maybe climbing upstairs is not easy for him, (he is) short of breath. Even this hard-breathing doesnt escape As ears. He stands behind door excitedly, waiting for the music to play. But its not clear whether B doesnt see the doorbell or (he) is too short to reach (it), he just raises his hand to knock on the door. A doesnt lose faith in B, but patiently waits for him to find the doorbell.
It is obvious that the difference between the two passages lies not so much in the choice of lexicon or complexity of the structure as in their planning of discourse and the compactness of encoding. Passage (50) is marked for its orality and limitation of on-line production. The speaker false-started three times in this short passage, repeated himself, stuttered once, and did some serious repair in the last sentence when he tried to describe the man’s reaction to the short guy’s failure to ring his doorbell. All that fumbling is laid bare for all to hear in oral narrative because when you speak you cannot destroy your earlier drafts (Halliday 1987: 69). The speaker has to keep communication flow, and a silence after each clause is very awkward in oral communication. Whereas passage (50) shows plainly the on-line planning, trial and error, and revision, passage (51) encodes the same events smoothly and succinctly and is free of all that fumbling because the preliminary scrap of the ‘script’ has been fixed before it is written.
The orality and informal style of oral narrative is further marked by the frequent use of exclamation, interjection or expletive markers in the CO group, such as哟in (52), 咳 in (53), 嗳in (54) below.
52) CO: S2
哟， 怎么 里面 有人 居然不给开门 啊！
Excl. how-come inside someone but not open-door Excl
Why! Someone is inside but does not opens the door!
53) CO: S10
咳，新装的 门铃 最终 还是没 派上用场！
Excl. new-install door-bell finally still not have use
Well, the new doorbell is not useful after all!
54) CO: S13
嗳，门 里边 有 声音 啊！
Excl. door inside have noise Excl
Ah, is someone inside?
The exclamation markers such as these express various kinds of feelings and emotion of the character or the speaker at the time of utterance such as disbelief (52), disappointment (53), surprise (54), certainty, satisfaction, anger, frustration, etc., or a combination of some of them. They grab the potential hearers attention by breaking the monotone of the narrative and make the oral discourse more colorful and interesting with their sharp rising and falling pitches, and moreover, they often reinforce the statement/question with their own emotional undertone. For example, the interjection咳in (53) clearly shows the characters disappointment and annoyance that nobody has rung his new doorbell. In addition, the disappointment and annoyance may well be the speakers as well since the statement is ambiguous as to its reference of origin, and the speaker generally empathizes with the central character of the story. Without the interjection marker, (53) would have been another simple statement in the storytelling. Whereas such exclamation markers are often employed by Chinese speakers, they very rarely occur in the written narrative because interjections show orality and informality, which run counter to the expectation of written language.
The second difference found between the two modalities at discourse level concerns the involvement of speakers and writers with the audience and their empathy with the main character of the narrative. Although no real listener is present before speakers in our narrative study, we have so far shown that the nature of orality in speaking and formality/literacy in writing is preserved between our spoken and written narratives. It seems that the role of hearer is an inherent part of the act of speaking, regardless of whether or not a hearer is present. In our study, speakers exhibit a higher degree of involvement with the potential audience than the writers. In producing English oral narratives, for example, many speakers try to get involved with their audience, though unseen and unknown, by using expressions such as you know, you hear, you see, etc. in their story-telling. Consider the following.
55) EO: S14
Someone came up the stairs which you cannot see, but you heard themcoming up the stairs.
56) EO: S17
Then you hear, um .. someone coming up the stairs again.
57) EO: S20
Um .. then you heard the light click on, then click off.
It appears that speakers not only interact with hearers but also try often to have them see and hear the story as if they were watching the video themselves. They try to recreate the story as it unfolds from an on-line video-watchers point of view by revealing what one sees, hears and infers. After all, when a hearer encounters the recorded story at a later point, s/he hears the speakers voice and all the suprasegmental details of the narrative discourse as well. It is much easier for a speaker than for a writer to engage the audience in a here-and-now situation. The speaker of (55), for example, tells the audience that someone comes upstairs, who is not seen but only heard. While half of the speakers used these interactive expressions to engage the audience in an on-going story, writers tell the story from a more static perspective as a story-teller but not as a fellow video-watcher with the audience. They follow the main story-line, often omitting audio details of the video. For example,
58) EW: S9
He takes a chair and waits for someone to ring it. A girl comes, but only ties her shoe.
59) EW: S1
He then pulls up a seat and opens the newspaper, waiting for a visitor to come and ring the bell. A woman comes up the stairs, he thinks shell ring. But doesnt. She goes up one more flight.
60) EW: S13
He then waits for somebody to come and ring the door bell so that he can hear it again. A person comes to the door and stops but he just ties his shoe.
When writers sometimes do mention the footsteps, noise, sounds, etc. in the video, they always used the third person singular reference instead of the first or second person pronoun. It is he (the man in the story) who hears those sounds, not I, you or we. For example,
61) EW: S2
He hears someone coming. So he closes his door and waits for them to press it.
62) EW: S16
Again he hears approaching footsteps.
Furthermore, speakers often show their involvement with the audience by using expressions such as I believe, I guess, I think, etc. to offer their personal opinion during story-telling, and reveal to the audience what they think may have happened in the story. For example,
63) EO: S14
And then the last guy was a mailman, I assume.
64) EO: S16
Then we can hear something that sounds like a motorcycle pull up, I guess.
65) EO: S19
And um so then the man s s s .. you know, he still sat in his chair and um .. Ibelieve somebody else came by.
The use of the first person reference in I assume, I guess, or I believe seems to indicate speakers’ willingness to inform the audience of their uncertainty or self-evaluation in storytelling, and these tentative expressions appear to imply that ‘what I have just said or what I am going to say may or may not describe the scene or the character precisely, but that is my understanding’. Without an audience in mind, speakers would not have employed the first person reference such as the ones in (63-65) above.
In general, speakers take the audience into consideration while they contemplate, infer and reach the tentative conclusion of what happens in the story because the audience is inherently part of the oral-storytelling process. This kind of involvement, however, never occurs in the English written narrative, where writers do not offer personal opinions but describe what happens