Factor scores and textual relations

The primary goal of this study is specification of the textual relations in English speech and writing, that is, the linguistic similarities and differences among English texts. To this point, six parameters of variation have been identified through a factor analysis and interpreted as underlying textual dimensions. In the present chapter, the similarities and differences among genres are considered with respect to each of these dimensions, and the overall relations among genres in speech and writing are specified by consideration of all dimensions simultaneously. Genres can be similar with respect to some dimensions but quite different with respect to others; the textual relations among genres are determined by the joint assessment of similarities and differences with respect to all dimensions.

Genres can be compared along each dimension by computing factor scores (see Section 5.5). To recapitulate, factor scores are computed by summing the frequency of each of the features on a factor, for each text; for example, the factor score of a text for Factor 2 might equal 23 past tense + 50 third person pronouns + 10 perfect aspect verbs + etc. The factor scores for each text can be averaged across all texts in a genre to compute a mean dimension score for the genre, and these mean dimension scores can be compared to specify the relations among genres.

Genres and text types

Genre categories are determined on the basis of external criteria relating to the speaker's purpose and topic; they are assigned on the basis of use rather than on the basis of form. It is also possible to consider groupings of texts that are derived on the basis of linguistic form. In other work (Biber forthcoming) I distinguish ‘genres’ from ‘text types’: genres characterize texts on the basis of external criteria, while text types represent groupings of texts that are similar in their linguistic form, irrespective of genre. For example, an academic article on Asian history represents formal, academic exposition in terms of the author's purpose, but its linguistic form might be narrative-like and more similar to some types of fiction than to scientific or engineering academic articles. The genre of such a text would be academic exposition, but its text type might be academic narrative.

Genres are not equally coherent in their linguistic characterizations. Some genres have several sub-classes which are quite different from one another; for example, academic prose includes engineering articles, political and historical analyses, and literary discussions. The linguistic form of texts in other genres is simply not highly constrained, and thus these genres permit a relatively wide range of variation; for example, the linguistic characteristics of face-to-face conversation in private academic settings, public social settings, and intimate settings are all different.

Introduction

A considerable body of research in the humanities and social sciences has dealt with the similarities and differences between speech and writing. Work in history, sociology, anthropology, psychology, education, comparative literature, and linguistics has described ways in which the choice between speech and writing is closely related to developments in other social institutions. For example, the development of widespread alphabetic literacy in ancient Greece was probably a catalyst for other social and intellectual developments there. Widespread literacy enabled a fuller understanding and participation by citizens in the workings of government, which might have promoted a democratic form of government in which citizens play a relatively active role. Literacy enabled a permanent, accurate record of ideas and the possibility of knowledge without a living ‘knower’. As such it probably aided in the transition from ‘myth’ to ‘history’ and the development of critical attitudes towards knowledge. Prior to literacy and a permanent record of beliefs and knowledge, a society can alter its beliefs and not be faced with the possibility of a contradiction; competing ideas which evolve slowly over generations will be accepted as equally factual when there is no contradictory record of earlier ideas. Written records, however, force us to acknowledge the contradictory ideas of earlier societies and thus to regard knowledge with a critical and somewhat skeptical attitude. For example, we know that earlier societies believed that the earth was flat, because these beliefs are permanently recorded in writing.

Overview of the study

The analysis presented here was undertaken to describe the relationship between speech and writing in English. Previous studies have offered a wide range of conclusions concerning this relationship; some studies conclude that speech and writing are not very different from a linguistic perspective, while others conclude that they are fundamentally different; some studies conclude that the differences between speech and writing are due to one set of features, while others focus on a different set of features. The present study sorts out these contradictory findings and arrives at an overall account of the textual relations in spoken and written English. To accomplish this task, the study analyzes the distribution of many functionally diverse linguistic features in many different types of spoken and written texts. This analysis shows that there is no single, absolute difference between speech and writing in English; rather there are several dimensions of variation, and particular types of speech and writing are more or less similar with respect to each dimension. In all, six dimensions of variation are identified here, which are interpreted in functional terms and labelled: Dimension 1: ‘Involved versus Informational Production’, Dimension 2: ‘Narrative versus Non-Narrative Concerns’, Dimension 3: ‘Explicit versus Situation-Dependent Reference’, Dimension 4: ‘Overt Expression of Persuasion’, Dimension 5: ‘Abstract versus Non-Abstract Information’, and Dimension 6: ‘On-Line Informational Elaboration’.

Summary of factor interpretation principles

In this chapter we turn to an interpretation of the factors presented in Chapter 5, to identify the construct or dimension underlying each factor. Recall that the factor analysis identifies groups of linguistic features that co-occur frequently in texts. The interpretation of the factors is based on the theoretical assumption that these co-occurrence patterns indicate an underlying communicative function shared by the features; that is, it is assumed that linguistic features co-occur frequently in texts because they are used for a shared set of communicative functions in those texts. The interpretation of each factor thus involves an assessment of the communicative function(s) most widely shared by the co-occurring features. Functional analyses of individual features in texts are crucially important in the interpretation process, since they provide the foundation for determining the function(s) underlying a set of features. In the present case, functional interpretations from previous research are summarized in Appendix II, and micro-analyses of features in particular texts are further discussed in Chapter 7. The interpretations presented here are based both on the findings of previous research and the analyses given in Chapter 7.

Table 5.4 (in Chapter 5), which summarizes the final factorial structure, is repeated here for convenience as Table 6.1. This table presents the important linguistic features comprising each factor.

Development of computer programs for grammatical analysis

One of the distinctive characteristics of the present study is inclusion of a large number of linguistic features representing the range of functional possibilities in English. Further, these features are counted in a large number of texts and genres, to exclude idiosyncratic variation and to insure inclusion of the range of situational and linguistic variation existing within speaking and writing in English.

The use of computerized text corpora and computer programs for the automatic identification of linguistic features made it possible to carry out a study of this scope. The programs, which are written in PL/1, use the untagged versions of the LOB and London–Lund corpora as input. In a tagged corpus, such as the Brown corpus, the words in a text are all marked, or ‘tagged’, for their grammatical category, greatly facilitating automatic syntactic analysis. A tagged version of the LOB corpus became available during the course of the present study, but it was not used because there is no comparable version of the London–Lund corpus (the spoken texts). That is, programs that took advantage of the grammatical tagging in the LOB corpus would identify features with a greater accuracy than could be identified in the London–Lund corpus, thus skewing the comparison of spoken and written genres. Therefore, the untagged versions of both corpora were used, and a single set of programs was developed for the analysis of both.

This study is based on both macroscopic and microscopic analyses of textual variation (see Section 4.1). Macroscopic analyses identify the dimensions of variation among texts and specify the overall relations among genres with respect to those dimensions. Microscopic analyses describe the functions of linguistic features in relation to the speech situations of individual texts. Linguistic features mark particular components of the situation, in addition to their functions as markers of relations within a text. In this chapter, I describe the salient components of the speech situation and provide a brief overview of the major communicative functions served by linguistic features. I then turn to the situational and functional differences between ‘typical’ speaking and writing and propose a framework for comparing more or less typical genres in terms of their situational characteristics.

Components of the speech situation

There are several studies that catalog the components of the speech situation, which provides the situational context for ‘speech events’. One of the earliest and most complete descriptions is presented in Hymes's (1974:53ff) components of speech, which include message form, message content, speaker, hearer, purposes, key, channels, and norms of interaction. This description is further elaborated by Duranti (1985). Fishman (1972) identifies three primary components of the situation of language use: (1) the participants and the relationship among them, (2) the topic, and (3) the setting. Halliday (1978) also distinguishes among three components of the communicative situation: (1) the type of social action (the ‘field’), (2) the role relationships (the ‘tenor’), and (3) the symbolic organization (or ‘mode’).

Macroscopic and microscopic approaches to textual variation

Within the broad framework of investigation into the psychological and sociological underpinnings of linguistic variation, researchers have investigated textual variation through macroscopic and microscopic analyses. Macroscopic analysis attempts to define the overall dimensions of variation in a language. Microscopic analysis, on the other hand, provides a detailed description of the communicative functions of particular linguistic features (e.g., clefts as markers of informational prominence, or first person pronouns as markers of personal involvement).

Much of the previous work analyzing linguistic variation in texts falls into the category of microscopic analysis. For instance, Schiffrin (1981) looks at the different functions of past tense and present tense forms in referring to past events in narrative. Aijmer (1986) and Stenström (1986) study the functions of actually and really respectively in conversational texts. Thompson (1983) studies the functions of detached participial clauses in descriptive texts. And Tannen (1982a) contrasts the level of imageability in written and spoken narratives to illustrate the use of oral strategies in written discourse. These and other studies are characterized by their detailed attention to the functions of specific features in representative texts.

Macroscopic analyses identify the overall parameters of linguistic variation within a given ‘domain’, e.g., spoken and written texts in English or the range of expository prose in English; they are based on the notions of textual dimension and textual relation.

The largest part of eastern Sumbanese ritual life consists of speeches conducted in a formal style which, following common usage, can be called a ‘ritual language’. This paper has two principal aims: (1) to locate ritual language performances within the general scheme of eastern Sumbanese culture, paying particular attention to the value this idiom holds for members of this society, and (2) to review some of its most prominent formal features, concentrating especially on attributes which may prove to be of comparative interest. The data on which this paper is based are drawn from my own fieldwork carried out in the domain of Rindi. However, since eastern Sumba is both culturally and linguistically a largely homogeneous area, and because, with minor variations, the same ritual language tradition is found throughout this region, in what follows I shall for the most part refer to eastern Sumba, mentioning Rindi only where context requires.

General remarks

Eastern Sumbanese ritual language differs from ordinary discourse in several fundamental respects. Like similar and related traditions in other parts of Indonesia, it is most succinctly described in the phrase ‘formal, formulaic, and parallelistic’ (Fox 1971:215). Parallelism is evidenced by a prescribed pairing of words to form dyadic sets which appear as components of longer expressions, thus producing an overall pattern of paired lines or couplets. This principle affects both the semantics and syntax of ritual speech. Thus paired elements, which may provisionally be classified as comprising synonyms, various types of antithesis or ‘synthetic’ combinations, are usually morphologically identical and grammatically equivalent, and appear in corresponding positions within parallel lines.

Introduction

The Ata Tana Ai (People of the Forest Land) inhabit the watershed and tributary valleys of Napun Geté, the largest riparian system of the Regency of Sikka in eastern Flores. Tana Ai is a region of well-watered forest land insulated from the neighbouring peoples of eastern Flores by high mountain ridges to the east and west. Within their valley the Ata Tana Ai have maintained a society and a ceremonial system that are quite distinct from those of the peoples of central Sikka to the west and Larantuka (Regency of East Flores) to the east.

Despite increasing pressures from both the state and the Catholic church, the Ata Tana Ai maintain their traditional religion and a ceremonial system that integrates clans and maternal houses into larger social and political entities. Tana Ai is divided into several ceremonial domains, called tana in the local language, each of which is led by a tana puan, who is ‘Source of the Domain’ or ‘Source of the Earth’. While they are generally comparable, the rituals and myths of the various tana of Tana Ai comprise distinct traditions. The largest tana of Tana Ai, and the one most important to the Ata Tana Ai themselves, is Tana Wai Brama. It is this central domain whose myths and ritual language are the subjects of this essay.