A Study of Narrative Structure and Cognitive Processes in English Language and Literature

In order to accurately analyze the characteristics of narrative structure and cognitive process in English language and literature, this paper aims to construct a corpus of English language and literature works and explore it by analyzing the changes in the data related to its works. In this paper, the texts of the corpus of English language and literature works are taken from the relevant English language and literature works on the Internet, the corpus is real, complete and reliable, and the selected English language and literature works are typical and representative.

In this paper, the source of corpus is mainly PDF format or Word format e-books, due to the uncertainty of the downloaded text, so it is necessary to manually sieve and screen the text to ensure that there is no garbled code in the text, no blank pages and other interference items. After downloading Adobe Acrobat software to edit PDF format files, the software can effectively retain the original file format while editing and converting PDF format text. This paper builds an observation corpus with a capacity of 123,696 words, which is in line with the principle of corpus construction and has a certain scale and representativeness, with greater research value and significance of discovery.

In addition, in order to compare the language changes in the corpus of English language and literature works, this paper further expands the data sources of English language and literature works, and collects the data related to English language and literature works on the global Internet as the reference corpus. The library capacity of the reference corpus totaled 675,148 words, and the overall library capacity is large, which is 5.46 times larger than the library capacity of the observation corpus under study. It is in line with the principles and standards of the reference corpus, which can basically meet the needs of the study and effectively support the research credibility.

Based on the collected English language literature works, this paper constructs an observation corpus and a reference corpus respectively, in which the observation corpus is mainly applied to explore the narrative structure and cognitive performance of English language literature works, and the reference corpus serves as the control of the observation corpus as a way to clarify the value connotation of English language literature works. Figure 1 shows the construction process of the corpus of English language literature works, which mainly contains five steps, namely, corpus collection, preprocessing, corpus production, corpus organization and corpus expansion [18].

Figure 1.

The construction process of the corpus of English language and literature

Since the corpus established in this paper is based on data related to English language and literature, the algorithms applicable to Chinese word segmentation are not applicable to English language and literature corpus segmentation. Based on this, this paper proposes an English language and literature corpus segmentation algorithm that combines word frequency tree and information theory. The details are as follows:

Firstly, the text data of English language and literature is sliced, and each unit after slicing is considered as a gram, and then the frequency of each gram is counted. According to the set threshold, the words that do not meet the word length requirement are filtered out. Due to the language-independent nature of N-gram, there is no need to carry out linguistic processing on the text data of English language and literature.Trie tree is a tree structure, which is a variant of hash tree.Trie tree utilizes prefixes to shorten the retrieval time, which can minimize the comparison of character strings, and it performs better especially in the scenarios of dynamically adding or modifying data.This paper constructs a gramT word frequency tree.

Mutual information and information entropy are both concepts in information theory.Mutual information calculates the degree of association between two objects.If X and Y are independent of each other, then X and Y do not provide any information to each other, and their mutual information is 0.N-Gram obtains text fragments with high frequency, but the high frequency of occurrence of a text fragment does not mean that the text fragment is a real word, it may be more than one words combined together in a phrase. In this paper, we use mutual information to measure the degree of cohesion of different text fragments [19]. Mutual information PMI(w) is calculated as: (1)PMI(w)=log2p(w)avg(w)$$PMI(w) = {\log _2}\frac{{p(w)}}{{avg(w)}}$$ (2)p(w)=f(w)num$$p(w) = \frac{{f(w)}}{{{\text{num}}}}$$

Where w = [w₁, …, w_n] represents the word composed of multiple characters, p(w) represents the probability of word w appearing in the large-scale corpus, f(w) represents the number of times word w appears in the large-scale corpus, num represents the number of words in the large-scale corpus, and avg(w) represents the average probability of different combinations of characters in the word.

Information entropy is used to judge the stability of the interrelationship between the left and right neighboring characters of English language and literature texts, the higher the entropy, the more informative the information and the higher the uncertainty, thus reflecting the richness of the collocation of the left and right neighboring characters. Information entropy is divided into left information entropy and right information entropy. Left information entropy is the degree of stability of a text fragment combined with characters on the left, and right information entropy is the degree of stability of a text fragment combined with characters on the right. Then: (3)En(w)=−∑wleft|rightp(wleft|right)log2p(wleft|right)$$En(w) = - \sum\limits_{wleft|\;right} p ({w_{\;left|\;right}}){\log _2}p({w_{\;left|\;right}})$$

where p(w_left|night) is the conditional probability of a neighboring character w_left|right to the left or right of candidate w when it appears. The computational expression for w_left|night is: (4)p(wleft|right)=N(wleft|right)N(w)$$p({w_{left|right}}) = \frac{{N({w_{left|right}})}}{{N(w)}}$$

Where N(w_left|night) is the probability that neighboring characters and candidate word w appear together, and N(w) is the probability that candidate word w appears alone.

In this paper, the above algorithm is used to realize the corpus segmentation of English language and literature works, and its purpose is to improve the mining ability of English language and literature works, and lay the foundation for analyzing the narrative structure of English language and literature works as well as the impact on the cognitive process of readers.

After completing the acquisition and segmentation of the corpus, it is also necessary to annotate the information of the corpus to be entered into the corpus. Labeling refers to the process of reprocessing the collected raw corpus and labeling it with some additional information to make it mature. It is necessary to label the corpus accurately and effectively. First of all, the annotation of a corpus can improve the efficiency of information extraction. After labeling the information, users can not only extract the information from the corpus itself, but also directly extract and absorb the information and knowledge outside the corpus. Secondly, after annotating the corpus information, all the corpus entered into the corpus will be changed from raw corpus to familiar corpus. The familiarized corpus can bring convenience to users, who can quickly understand the collected corpus and use search tools to retrieve it.

The labeling information of this corpus mainly appears in the labeling of the cooked corpus and the statistical database, in which the statistical database is mainly labeled with statistical information and categories. The labeling of the cooked corpus is the most important part of the information labeling of this corpus, including the basic information labeling, additional information labeling, lexical description labeling and usage comparison labeling in four parts, which are specifically subdivided into the word form labeling, pronunciation labeling, word level labeling, lexical relationship labeling, high-frequency inverse-order words, easy to confuse inverse-order words labeling, structural labeling, lexicality labeling, paraphrase labeling, and discursive labeling and so on. The corpus is manually labeled, checked, and proofread again after the labeling is completed. In the process of labeling the corpus, the principles of reasonableness, flexibility, clarity, and lack of labeling are strictly followed.

In English literature, narrative structure is the cornerstone that supports the framework of the whole story, and the basic elements that constitute this framework are plot, character, and setting. These three are interwoven with each other and together constitute the rich connotation of literary narrative structure. 1)

Plot is the core of literary narrative structure. It refers to the development process of a story, including the cause, passage and result of events. The setting of the plot needs to be skillful and reasonable, able to arouse the reader’s interest, promote the development of the story, and reach the peak of emotion in the climax. A compelling plot can enable readers to be completely immersed in the story and empathize strongly with the characters.

In the narrative structure of English language literature, plot, characters, and environment are interdependent and mutually influential. The development of the plot requires the participation and promotion of the characters, while their character and behavior are influenced by the environment. At the same time, the environment also provides rich materials and inspiration for the plot and characters, making the whole narrative structure more complete and three-dimensional.

English language literature presents a temporal and spatial background of ambiguous literary creation through mythological stories and social reality stories, innovating narrative expression through traditional narrative structures. On the one hand, English language literature builds up the framework of the desired society through words, helping readers to re-cognize themselves and perceive the society, so as to provide effective reference programs for the problems existing in the real society. On the other hand, it makes the relationship between the readers and the authors gradually harmonious, until the readers also become part of the authors, and finally brings a new turn for the future development of English language and literature. 1)

English Language and Literature can help readers to re-cognize themselves and lead them to recognize the social reality. That is to say, while focusing on the progress and development of oneself, it should also reflect on the current development of society, address problems, and provide effective solutions.

The chi-square (CHI) method is employed to measure the degree of association between an entry and its category by calculating the difference between the data and a hypothetical distribution. If the chi-square value is larger, the more related the lexeme is to the category and the more the lexeme embodies the category, and if the chi-square value is smaller, the less related the two are and the less the lexeme embodies the category [20]. Table 1 illustrates the connection between feature words and categories.

The computation method based on the feature word t is: (5)E11=(A+BN)×(A+D)$${E_{11}} = \left( {\frac{{A + B}}{N}} \right) \times (A + D)$$ (6)E12=(A+BN)×(B+D)$${E_{12}} = \left( {\frac{{A + B}}{N}} \right) \times (B + D)$$ (7)E21=(C+DN)×(A+C)$${E_{21}} = \left( {\frac{{C + D}}{N}} \right) \times (A + C)$$ (8)E22=(C+DN)×(B+D)$${E_{22}} = \left( {\frac{{C + D}}{N}} \right) \times (B + D)$$

The correlation between feature word t and category is calculated as according to the chi-square detection formula: (9)D11=(A−E11)2E11$${D_{11}} = \frac{{{{(A - {E_{11}})}^2}}}{{{E_{11}}}}$$

Similarly to find D₁₂, D₂₁, D₂₂, substitute and solve for the χ²-values of the feature words t and c_i as: (10)χ2(t,ci)=N(AD−BC)2(A+B)(A+C)(B+C)(C+D)$${\chi ^2}(t,{c_i}) = \frac{{N{{(AD - BC)}^2}}}{{(A + B)(A + C)(B + C)(C + D)}}$$

From the above formula, it can be seen that when word t has lower correlation with category c_i, the closer the value of χ² is to 0. When feature word t has stronger correlation with category c_i, the larger the value of χ² is.

The steps for selecting the feature words by the chi-square statistic are as follows: 1)

List all words i_i, 1, i_i,2, i_i,3 ⋯, t_i,j in category c_i.

Based on the feature words in the English language and literature corpus obtained by chi-square statistics, this paper adopts the improved dichotomous K-Means clustering algorithm to obtain the narrative keywords in the corpus of English language and literature works. The main idea is to find a center of mass randomly, and then according to the spatial distance to find the more dispersed data points with the largest SSE difference as the initial center of mass, to initially divide the entire dataset space. Then use the spatial distance of the discrete distribution of spatial data as the basis of clustering for binary classification of clusters, which can better improve the running speed of the algorithm [21]. Its specific application steps are as follows:

Let there be n point to be clustered in data set X as X={x1,x2,x3,⋯,xn}$$X = \left\{ {{x_1},{x_2},{x_3}, \cdots ,{x_n}} \right\}$$, we need to cluster it into k clusters, Mi(j)$$M_i^{(j)}$$ denotes the ith data point in cluster j(j = 1, 2, ⋯, k), Z_j is the center of mass of cluster i, n(j) denotes the number of data objects in cluster i, firstly, we classify all the data points in the same cluster, and then we proceed with the following steps: 1)

At the beginning of splitting the whole data set we find a random data point as the center of mass Z_j, then we first calculate the distance D(Mi(j),Zj)j∈[1,k]$$D{(M_i^{(j)},{Z_j})_j} \in [1,k]$$ of each data point to the center of mass Z_j to find the data point with the farthest distance from the center of mass Z_j noted as K_j, then: (11)Zj=1n(j)∑i=1n(j)Mi(j)(j∈[1,k])$${Z_j} = \frac{1}{{n(j)}}\sum\limits_{i = 1}^{n(j)} {M_i^{(j)}} (j \in [1,k])$$ (12)D(Mi(j),Zj)=ǁMi(j)−Zjǁ$$D(M_i^{(j)},{Z_j}) = ||M_i^{(j)} - {Z_j}||$$ (13)D(Kj,Zj)=max{D(Mi(j),Zj),i∈[1,n(j)],j∈[1,k]}$$D({K_j},{Z_j}) = \max \left\{ {D(M_i^{(j)},{Z_j}),i \in [1,n(j)],j \in [1,k]} \right\}$$

Narrative is a form of storytelling, storytelling is the oldest form of literature and the origin of storytelling can be traced back to the earliest stages of the formation of human civilization, thus narrative is a tradition. This means that in the internal context of literary development, the novel is an extended extension of the narrative tradition. The historical and cultural context in which English language literature is written determines the plurality of narrative traditions it inherits. Based on the corpus of English language literature works constructed in this paper, the narrative forms of English language literature works during the period of 1900~2000 are counted, and the trend of narrative forms of English language literature is obtained as shown in Figure 2.

Figure 2.

The change of narrative form

As can be seen from the figure, the narrative forms of English language literature during the period of 1900~2000 are mainly manifested in the three types of translation, adaptation and utilization, which together contribute to the change of narrative forms of English language literature. Among them, the narrative form of translation showed a trend of growth and then decline between 1900 and 2000, and the number of English language literature works increased by 270 between 1900 and 1945, with an overall growth rate of 94.41%. However, between 1945 and 2000, English language literature showed a decreasing trend, and in 2000, the number of English language literature in the form of translated narrative decreased by 469 compared with that of 1945, with an overall decrease of 84.35%. Translated narrative form allows oral narrative texts to be transformed into written texts, which became the most dominant narrative form of English language literature before 1945. And after that, the emergence of two narrative forms of adaptation and utilization led to the gradual demise of the translated narrative form. The two narrative forms of adaptation and utilization are more in line with the expression of English language literature, which can be based on different degrees of oral narrative to adapt to the contemporary needs and concepts of literary expression, and combined with certain adaptations and unique techniques to reproduce the stories and other elements of the oral narrative tradition in the form of fiction. As a result, the trend of adapting and utilizing these two narrative forms between 1900 and 2000 belongs to a continuous growth, and its overall growth rate reaches an astonishing 143.14% and 240.38%, respectively. Moreover, English literary works in the form of utilizing narratives are more capable of expressing the connotation of oral narratives, which compel writers to take the background of the times as the basis, and seek different ways of expression for their cultural adequacy, and then realize the metaphorical or symbolic use of traditional culture, which provides a new paradigm for the innovation of the narrative form of the English language and literature works.

Novel is the art of narrative, and narrative is inseparable from the main body of the narrative, the main body of the narrative refers to the narrator of the story content in the narrative process and the expression of the narrative intent. For English literary works, the narrator of the text has absolute authority, and the subjective consciousness of the author and characters implied in the work is often strictly under the control of the narrator. The skillful use of multiple narrative perspectives in literary creation can give different groups of people different narrative expressions, thus building a fascinating world of English literature. As one of the most complex and meaningful issues in narrative literature, perspective is a very important tool for conveying thematic meaning, and the author’s choice of narrative perspective determines how much information is output. Statistics on the literary narrative perspectives in the corpus of English literature works show the distribution of different types of narrative perspectives, as shown in Figure 3.

Figure 3.

Different types of narrative perspectives

Narrative perspectives in English language literature are mainly divided into three categories: inward-focused, outward-focused, and zero-focused. Inward-focused narrative is a perspective in which the narrator is equal to the characters in the work. This kind of work adopts the perspective of the character from the beginning to the end, and the narrator is the main character of the text, depicting the external world in the eyes of the main character with the first-person inward focus, and showing the main character’s inner thoughts and struggles. The externally-focused narrative point of view refers to the point of view of the narrator who is less than the character of the work. This kind of narrative perspective focuses on outlining the character’s words and deeds from the outside and does not participate in the inner activities of the protagonist, so the objective and calm narrative style is its specialty. At the same time, due to the limited perspective, the author’s description is not focused on exposing facts or truths, but on metaphysical thinking about the philosophical discourse of the subject. Zero-focus is an omniscient narrative perspective, which believes that the narrator is greater than the characters in the work, and the shifting and multiple perspectives in internal focus are also part of the narrative mode in which the narrator knows more than the characters, so any change in the angle of observation is categorized as a zero-focus narrative perspective.

During the period from 1900 to 2000, English language literature adopts the inward-focused narrative point of view relatively less, and its mean percentage is about 25.41%, which is 15.4 and 8.37 percentage points lower than that of the outward-focused and zero-focused narrative point of view. Although literary creation is accomplished through the first-person narrative perspective, it also provides a new artistic means to express the spiritual and emotional world of human beings, which not only accomplishes the multiplicity of meanings and narrative tension of literary works, but also enhances their mellow and timeless aesthetic meaning. However, relatively speaking, out-of-focus and zero-focus narrative perspectives utilize time-space intervals in the plot setting, which requires readers to pay more attention to understand the author’s hidden meaning through the text, as well as to understand the ineffable things that are hidden behind the life.

The innovation of narrative structure in English language literature mainly lies in the change of non-linear narrative structure, which is the key part of English language literature. Literary authors are skilled at creating complex narrative structures that aim to express the multiple profound subjects of their works and provide readers with a more diverse reading experience. In order to analyze the change of narrative structure in English language literature, this paper is based on the self-constructed corpus, combined with the cardinality statistics of its feature words, and utilizes the improved quadratic K-Means clustering algorithm to cluster and analyze the keywords of narrative structure. Figure 4 shows the clustering map of keywords in the narrative structure of English language literature.

Figure 4.

Narrative structure keywords cluster map

The cluster analysis mapping shows that the number of network nodes and the number of connecting lines are 528 and 349, respectively, with an overall network density of 0.083 and Q and S values of 0.849 and 0.956, respectively. Each parameter meets the criteria of the clustering map, which indicates that the results of this clustering are more reasonable and have a high degree of confidence. In this study, a total of eight clusters were formed for the narrative structure of English language and literature, and an inductive summary of the eight clusters revealed that the narrative structure of English language and literature is centered around three main themes, namely, bilinear narrative (#0 multifaceted mapping, #1 temporal measures of disarrangement), nested structure (#2 hierarchical relationships, #3 narrative embeddedness, and #4 multilayered structure), and collage structure (#5 narrative text, #6 narrative situation, and #7 (narrative space).

The two-line narrative structure mainly arranges the two stories of a literary work together, and the narrator does not have time-turning padding when telling the story, simply presenting the two stories to the readers. After reading, the reader will understand the original story. Such a two-line narrative pattern makes the story richer and more comprehensive, reflecting one of the propositions of life that the author wants to explore.

Nested structures are an important structural form in English language literature, in which there are references by the storyteller to a story that is about to be told, that is, an internal narrative structure that is about to unfold the narrative nested within an external narrative structure. These internal structures can be interconnected or independent of each other, but they are not directly related to the external narrative structure. In the nested structure of Monroe’s narrative work, the internal narrative structure is actually a supplement to the external narrative structure, and the two are combined together to form a complete story.

Collage structure refers to a narrative structure in which multiple narrative modules, including modules from different narrative texts or different narrative situations, are collaged together. This narrative mode was occasionally adopted by some genre writers in the modernist phase and was consciously materialized in the postmodernist literary period, and collage literature also embodies the narrative strategy of an era with its real existence. When the reader reads the whole novel, they will realize that the originally fragmented parts are actually all indispensable parts of the narrative structure.

Before the cognitive process-guided analysis of English language and literature, it is necessary to analyze the overall characteristics of its language and elaborate the lexical and morphological characteristics of English language and literature at the macro level. Based on the corpus of English language literature and the reference library constructed in the previous section, the overall stylistic style of English language literature is analyzed in terms of such indicators as class symbol to form symbol ratio (STTR), mean word length (MWL), standard deviation of word length (WL.std), and percentage of ≤5-letter words (≤5%). Table 2 shows the overall characteristics of the English language and literature corpus.

As can be seen from the table, the STTR value of the self-constructed corpus of English language and literature works (53.64) is much lower than the STTR value of the reference library (60.87). This indicates that the text data of English language and literature works have weak lexical diversity, and their word changes are all lower, i.e., there are more word repetitions in English language and literature works texts compared to the reference library. The average word length of the English language texts is 1.57% higher than the reference library, and it is necessary to test whether the difference between the data is significant at the statistical level. The t-test yields its t=9.542*** (p=0.002<0.01), indicating that the difference between the data is statistically significant. The comparison results reflect that the average word length data of English language and literature is larger than the relevant data of the reference library, and the vocabulary difficulty is higher than that of the reference library, and the tendency to use longer vocabulary in English language and literature can be seen from the level of average word length.

From the point of view of the percentage of words with less than 5 letters, the English Language and Literature corpus (42.64%) is slightly lower than the reference corpus (43.06%), but the difference in their data is not statistically significant (Chi-Square=0.5732/p=0.453, LLR=0.316/p=0.532). This data can indicate the lexical difficulty of the text, if not tested for significance, then the lexical difficulty of the English language and literature works is lower than the reference library, but the difference between the two data is not significant, so the original hypothesis should be rejected. That is to say, the lexical difficulty of the corpus of English language and literature is not lower than that of the reference corpus, and the lexical difficulty of the corpus has to be examined from multiple perspectives.

Narrative structure in English language literature causes changes in readers’ cognitive behavior, and the frequency and percentage of cognitive behavior and chapter behavior realizations in English language literature corpora are counted to understand the cognitive behavior guidance process in English language literature. Table 3 shows the statistical results of frequency and percentage of realization of cognitive behavior and chapter behavior in different corpora.

As can be seen from the table, the frequency of guiding readers to implement chapter behaviors in English language literature is 1088 times and 2220 times in the observational and reference libraries, respectively, and their behaviors are mainly realized in exofocused and zero-focused narrative structures in the observational and reference libraries of English language literature, with the frequency of (43.11%, 36.58%) and (36.71%, 33.20%), while the application of in-focus narrative structure is relatively less. This is more consistent with the optimization results of narrative perspectives of English language literature given in the previous section, that externally-focused and zero-focused narrative structures can better introduce readers to the literary works and make them fully appreciate the author’s creative state of mind. In terms of guiding readers to perform cognitive behaviors, the frequency of guiding readers to perform cognitive behaviors in the observation and reference pools is 712 times and 1134 times respectively. Proportionally, the main part of English language literature focuses on the use of the narrative form of accommodation, whose proportion in the corpus of English language literature reaches 45.79%, and the proportion of application in the reference library reaches 50.09%. Comparatively speaking, the frequency of using the translated narrative form is low, with 115 and 223 instances in the two corpora respectively. This fully demonstrates that the cognitive process of English language and literature is dominated by out-focused and zero-focused narrative structures, combined with adapted narrative forms to mobilize readers’ thoughts and cognition, so that they can fully perceive the internal emotional expression of English language and literature works, and thus make corresponding cognitive behaviors.