PROCESS VARIATION IN TWO DISCOURSE GENRES IN RUSSIAN
by
Elena Maximova
A Master’s thesis submitted in partial fulfillment of the requirements for the degree of
Master of Science
(Clinical Linguistics)
at the Joint European Erasmus Mundus Master’s Programme in Clinical Linguistics (EMCL)
UNIVERSITY OF POTSDAM
November, 2010
PROCESS VARIATION IN TWO DISCOURSE GENRES IN RUSSIAN
Elena Maximova
Under the supervision of Professor Elizabeth Armstrong at Edith Cowan University, Australia
Second supervisor: Professor Irina Sekerina at the City University of New York, the USA
ABSTRACT
There are numerous studies written about discourse. However, the discourse analysis is not widely applied to clinical populations. In order to test the communication skills of aphasic speakers, non-brain-damaged speakers should be first examined. The current study addresses the variation of process production in the discourse of unimpaired Russian speakers and investigates the impact of the discourse genre on the pattern of the relative frequency of different kinds of processes. As to attain these goals, ten adult non-brain-damaged speakers of Russian were asked to produce speech samples in two different genres of recount and opinion. Transitivity analysis based on Systemic Functional Linguistic approach proposed by Halliday and Matthiessen (2004) was further used to analyse the resultant discourse.
The investigated variation of process types in the genres of recount and opinion appeared to have certain patterns. Thus, material processes were of the highest frequency in both genres. Relational and mental process types were less frequently used in these genres.
Verbal and behavioural processes had the smallest proportion of usage in the two genres.
Such a pattern of process variation within two genres is the reflection and characterization of each of them. Thus, the genre of recount is defined as the listing of events (presupposes the high usage of material processes), while the opinion genre includes the elements of mental and verbal argumentation (presupposes the high usage of mental and relational processes).
Making a comparison of each process type across two genres, the following pattern could be
observed: (1) the higher frequency of material and behavioural process types in the genre of recount than in the opinion genre; (2) the higher frequency of relational, mental, and verbal processes in the genre of opinion compared to the recount genre. Again, such a pattern of process variation across two genres is the reflection and characterization of each of them. So, the current study suggests that transitivity analysis can be a useful tool in investigation of process variation in discourse of unimpaired speakers before applying the similar analysis to clinical population.
ACKNOWLEDGMENTS
I am heartily thankful to my supervisor, Professor Elizabeth Armstrong, for giving me the guidance, encouragement and the inspiration, for supporting me with valuable advice and helpful comments through my working on the project. I also owe my deepest gratitude to my second supervisor, Professor Irina Sekerina, for giving me indispensable assistance, meaningful explanations, for availability, constructive criticism of the drafts. Especially I would like to thank my parents for their invaluable help with data collection and Sean Veld for his help with revising. This thesis would not be possible without the participants who kindly agreed to collaborate in this study.
TABLE OF CONTENTS
ABSTRACT...3
ACKNOWLEDGMENTS...5
LIST OF TABLES...8
LIST OF FIGURES...9
1 Introduction ………10
1.1 A brief introduction to discourse analysis ………...10
1.2 Systemic Functional Linguistics (SFL)………...……….10
1.2.1 A brief characterization of Systemic Functional Linguistics...12
1.2.2 System of transitivity and process types...15
1.2.3 Application of SFL...20
1.2.4 Challenges of applying SFL to Russian...23
1.3 Genres of recount and opinion...25
1.4 Hypothesis and predictions...29
1.4.1 Process production variation within genres...29
1.4.2 Process production variation across genres...30
2 Method...32
2.1 Participants...32
2.2 Discourse elicitation procedure...33
2.3 Data transcription...34
2.4 Data analysis...38
3 Results...39
3.1 Sample length...39
3.2 Statistics...39
3.3 The pattern of process use in the genre of recount...40
3.4 The pattern of process use in the genre of opinion...43
3.5 Process use across two genres of recount and opinion...46
3.6 Individual performance...49
4 Discussion and conclusions...52
4.1 Overview of findings on process production variation within genres...52
4.2 Overview of findings on process production variation across genres...55
4.3 Clinical applications...56
4.4 Conclusions...56
REFERENCES...59
APPENDIX 1 Examples of process types...63
APPENDIX 2 An analysis of a sample from the genre of recount (Hasan, 1985)...66
APPENDIX 3 An example of a speech sample analysis...69
LIST OF TABLES
1.1 Summary of the main process categories and their discourse functions (Halliday, 1994) with examples...…..17 2.1 Participant characteristics...33 3.1 Number of clauses produced by each participant in each genre...40
LIST OF FIGURES
1.1 Stratification...………..….15
3.1 Percentage of process types in the genre of recount...41
3.2 Percentage of process types in the genre of opinion...43
3.3 Percentage of process types across two genres of recount and opinion...47
Chapter 1 Introduction
When working with people with aphasia, clinicians need to evaluate communication abilities of their clients. For this purpose, clinicians use aphasia tests which are relatively clear-cut.
However, the evaluation of the aphasic speakers' abilities to perform in everyday discourse is a much more complex task which requires more interpretation. Additional complexity is a wide range of communication styles across multiple domains.
Through this study we aim to learn more about the application of transitivity analysis to the production of non-brain-damaged speakers of Russian in two different discourse genres of recount and opinion. We study unimpaired language users to establish how much variation in their performance can be expected before applying the same clinical task to brain-damaged speakers. Another important goal is to determine the impact of discourse genre on the variation of normal speakers' performance.
1.1 A brief introduction to discourse analysis
The broad definition of discourse is "language in use". Speakers use discourse as one of the primary resources for communication in everyday life. Discourse can be of a number of different sub-types, for example, recount, procedure, opinion, exposition. The Traditional
formal linguistic analysis is focused on the language at the level of the single word and sentence. The Discourse analysis has broadened this focus by addressing the way the speakers produce oral and written speech that is contextually appropriate and meaningful.
Word and sentence analysis is included in discourse analysis, however, in the sense in which they contribute to the overall meaning of discourse.
Over the past twenty years, numerous studies have been written about discourse.
Throughout this, the application of discourse analysis to clinical purposes is not widely done and still needs to be further developed as it is of value in the assessment and treatment of aphasia. Discourse evaluation is a complex task and requires more interpretation.
Specifically, there is an opportunity to learn more about the communication skills of fluent and non-fluent aphasic speakers. However, primarily, non-brain-damaged speakers should be examined in order to address the variation in their discourse to further investigation of discourse use by aphasic speakers.
The current study aims to examine types of processes produced by non-brain- damaged speakers of Russian in two different discourse genres (recount and opinion).
According to Halliday & Matthiessen (2004), five process types can be distinguished:
material, behavioural, verbal, mental, and relational. These processes allow speakers to express doing, happening, saying, thinking and being. Processes are realized by the main verb or verb phrase within a clause. In this study, Transitivity Analysis will be used to address its potential for evaluating variation of process production in fluent and non-fluent aphasic speakers when undertaking a similar analysis. The investigation of process variation is of great significance for examination of aphasic speech since verbs which realize processes are reported to be impaired in aphasic language.
Two main goals are specifically set in the current study of unimpaired population. The first goal is to establish how much variation in process types can normally be expected before
further application of the same clinical task with brain-damaged speakers. Before determining patterns as "pathological", more information is required to establish what is "typical" for these genres. The second goal is to determine the impact of the discourse genre on the pattern of the relative frequency of the different kinds of processes in unimpaired speakers. This information can be further used to more reliable evaluation of the usage of process types by impaired speakers as well as to determine whether a transitivity analysis is a reliable and suitable clinical measure.
1.2 Systematic Functional Linguistics
1.2.1 A brief characterization of Systemic Functional Linguistics
Systemic Functional Linguistics (SFL) is a model of grammar developed by Michael Halliday in the 1960s. It is focused on language in its social context. This model is viewed in terms of discourse and situation in which interlocutors have certain communicative goals.
Semantic and syntactic tools are provided to speakers through different dimensions. Bache (2010) explains that SFL is called "systemic" because language is described in the paradigmatic dimension which implies "choice relations". Hjelmslev (1943) defines such relations as "either-or" relations. An example of paradigmatic relation can be the choice between present and past in a sentence like He likes/liked her a lot ("either-or" relation: either likes or liked). In contrast to Systemic Functional Linguistics, Chomsky's generative linguistics, typological linguistics, and cognitive linguistics mainly focus on syntagmatic relations which can be defined as "horizontal", "linear", or "both-and" relations. Such relations form chains or sequences, as in the direct object noun phrase "the book" in I gave her the book where the relation between article and head noun is syntagmatic ("both-and"
relations: both the and book). Syntagmatic relations are often referred to as "syntactic".
Paradigmatic relations in SFL are formed by "systems" (a system of definiteness, a system of tense, etc.) which are often interrelated. Interacting with each other, the systems form complex "networks". For instance, the system of definiteness interacts with the number system in choice of the singular for the indefinite article a/an, and the singular or the plural for the definite article the. Thus, a language is regarded as a system of systems and consists of system networks (Halliday, 1981).
The second part of the name of Systemic Functional Linguistics is "functional".
Bache (2010) explains it as language to be in its social and cultural contexts. Thus, language is always in context, and context is always in language. Language is defined as purposeful social behaviour with communicative goals. Halliday (1978) considers language as a social semiotic system. He sets a direct relationship between the grammatical systems of language and the functional aspects associated with language and its use in context.
Within this framework, so-called "metafunctions" are realization of the relationship between systems and functions. Metafunctions organize language resources for the purpose of certain communicative goals (Martin, 1996). There are three metafunctions within SFL framework: (1) the ideational (which is subdivided into the experiential and the logical); (2) the interpersonal; and (3) the textual. According to Halliday (1978), the ideational metafunction is a function for organizing, understanding, and expressing our perceptions of the world and of our own consciousness. The interpersonal metafunction is a function of creating interpersonal relationships. The textual metafunction serves to organize information.
As Martin puts it, ideational resources have a strong representational function, interpersonal resources have a distinct social function, and textual resources enables the semiotic link between ideational and interpersonal resources.
The metafunctions serve as a link between grammatical systems and context. The interpersonal metafunction is expressed through the grammatical system of mood which
structures the clause in terms of mood and residue. The textual metafunction is expressed through the grammatical system of theme which structures the clause in terms of Theme and Rheme. The ideational metafunction (specifically experiential) is expressed through the grammatical system of transitivity which structures the clause in terms of participant and process. Thus, the ideational system at clause rank is transitivity. The system of transitivity is concerned with construing one particular domain of our experience which is the flux of
"goings-on": configurations of a process (material, relational, mental, verbal, behavioural), the participants involved in it (Actor, Goal; Senser, Phenomenon; and so on), and the circumstances attendant on it (Clause, Location, Manner, and so on). The transitivity system of a language construes experience into a small set of domains of meaning which differ according to the process itself and the nature of the participants involved in it. Thus, Systemic Functional Linguistics considers the transitivity system as a system which captures different semantic elements of the clause and relates them to overall discourse patterns. It is described as the grammatical resource for how we use language to describe our experience of the world.
One more characteristic of SFL is "stratification" of language. In contrast to traditional grammars, which distinguish such levels as phonology, semantics, and syntax, in SFL language is viewed as complex stratified system. In this system, language is structured into "expression" (sound) and "content" (meaning). The "expression" aspect is further stratified into phonetics and phonology, while the "content" aspect is further stratified into lexicogrammar and semantics. The lexicogrammar stratum is viewed as a close link between vocabulary items and syntactic structure. The stratum of semantics is defined as the semantics of discourse, not of individual clauses or words. Moreover, language is embedded in context, as SFL proposes the close link between language and context. The stratum of context is described in terms of Field, Tenor, and Mode with their strong relation to the three
metafunctions: ideational, interpersonal, and textual respectively. In Figure 1.2 (Halliday and Matthiessen, 2004), stratification is represented by means of concentric circles.
Fig. 1.1. Stratification.
Therefore, a model of grammar by SFL is represented as a complex system of interrelated components. Thus, language is closely linked to context. In this framework, the link between grammatical systems and functions of language is realized through three metafunctions (ideational, interpersonal, textual). Grammatical systems, in turn, are linked to the three categories (Field, Tenor, Mode) which are viewed in terms of the organization of the context of a situation. Additionally, language is stratified into levels of "expression"
(phonetics and phonology), "content" (lexicogrammar and semantics), and "context".
1.2.2 System of transitivity and processes types
Systemic Functional Linguistics proposes that the system of transitivity represents ideational content (the semantic content of language). The system of transitivity is therefore focused on
the clause as its unit of analysis. A clause is composed of two essential elements: the Process and the Participant(s) involved in it. Processes allow participants to express doing, thinking, happening, saying, being, or, "goings on" and "doings" that constitute life experience. A speaker may express (either externally or internally) different experiences via different types of processes. Information such as interests, daily experiences, opinions, likes and dislikes can be expressed by a range of processes. Processes are realised by the main verb or verb phrase within a clause, for instance, "he misses her", "he was missing her". Halliday (1994) distinguishes the following process types: material, relational, mental, verbal, and behavioural. The process types are described in detail below and are summarised in Table 1.1 (Appendix 1 contains further examples of process types).
Material processes
In the system of transitivity, material processes are the largest and the most diverse type of processes (Thompson, 2004). According to Halliday and Matthiessen (2004), they are processes of "doing" and "happening". Material processes describe actions in the material worlds in terms of physical actions and happenings (e.g., "run", "throw") and changes unfolding through time and space (e.g., "bake", "break"). The material type of processes is characterized by an expense of energy, or effort. Material processes usually present an Actor (the "doer" of the action) and a Goal (the "participant" whom the action aimed at). In an example presented by Motta-Roth and Nascimento (2009), in a clause "I was cooking dinner", there is a main Process which is material ("cooking"), an Actor who performs the action ("I") in relation to a Goal ("dinner"). Prototypically, material processes deal with concrete doings and happenings since their function is to construe the world in terms of
Table 1.1. Summary of the main process categories and their discourse functions (Halliday, 1994) with examples.
Process type Definition Function Examples
Material
Relational
Mental
Verbal
Behavioural
Processes of
"doing" and
"happening"
Processes of
"being" and
"existing"
Processes of sensing - "feeling",
"thinking" or
"perceiving"
Processes of
"saying"
Processes of
"psychological" or
"physiological"
behaviour
Recounting actions, events, and happenings
Description, evaluation/opinion,
categorization, connections of discourse,
participants Evaluation/opinion, reflection, conveying
feelings, reporting experiences Reporting conversation,
metaphorical usage
Reporting of physiological or psychological "actions"
John arrived She stirred the coffee It happened one night
Sarah is wise Peter has a piano
Today is forever
Some like it hot I hear them playing Remember that face
She explained She said her piece She speaks Russian
He snores loudly They sleep tight Betty cried a lot
physical experience (Halliday, 1994). However, they can also be abstract (e.g., "the fight dissolved", "she resigned") and can take active or passive voice.
Relational processes
Relational processes are processes that describe the state of being (usually realized by the verbs "be" and "have") and existing (usually realized by the verb "be"). The main function of relational processes is to relate one aspect of our experience to another in different ways.
According to Halliday (1994), relational processes relate the participant to his or her identity and/or descriptors by means of two different modes: attribution and identification. Relational attributive processes establish a relation of class membership between two Participants by signaling that an entity (a Participant) has "some class ascribed or attributed to it" (Halliday
& Matthiessen, 2004). For example, in a clause "the weather is great", the relational process
"is" is attributive because a speaker assigns quality to the weather. Relational identifying processes establish a relationship between two entities in which one of them ("Identifier") is used to identify another ("Identified") (Halliday & Matthiessen). For instance, in a clause
"Tom is a doctor", the relational process "is" is identifying since one entity (Identifier
"doctor") identifies the other (Identified "Tom").
Mental processes
Mental processes are processes of sensing which allow speakers to express goings-on in their mental world. According to Martin, Matthiessen, and Painter (1997), these processes have three main modes of sensing: perception (perception through the five senses), cognition (thinking), and affection (feeling). They represent processes of cognition which are not directly observable and take place in the mind. Mental processes do not indicate any kind of action since they do not have observable realization. Martinec (2000) says that in language mental processes have no direct realization. They are bidirectional which means that mental processes can support the active voice being reversible, as in a pair: "I like it-it pleases me".
In the case of mental processes, the Participant who is a conscious doer of the process is the Senser, and the Phenomenon is the mental concept in the mind of the Senser. For example, in a clause "I think", "I" is the Senser, and "think" is the Phenomenon.
Verbal processes
Verbal processes are processes of saying, and their main function is to provide a speaker with resources to enable a dialogue. Speakers can use verbal processes to report (as in: "She said she loves him") or quote (as in: "She said "I love you""). Halliday (1994) describes four participant roles associated with the verbal process type: the Sayer (a doer of the process), the Receiver (the one to whom the verbalization is directed), the Target (the one who gets a verbal action about), the Verbiage (the name for the verbalization itself). Even though the Sayer is usually a human being, it can also be anything which puts out a sign, as in "The sign tells you where to turn" or "The siren screams loudly".
Behavioural processes
Behavioural processes are processes of physiological or psychological behaviour. Halliday and Matthiessen (2004) give such examples of behavioural processes as breathing, smiling, coughing, seeing, etc. Behavioural processes can be divided into intro-active (with one participant) and interactive (with two or more participants). Intro-active processes can be further divided into ranged and non-ranged. Ranged behavioural processes are processes of grooming (for example, combing, dressing, shaving, etc.). Non-ranged processes can be further classified into affective (those which are associated with positive or negative affect, e.g., crying, giggling, smiling, etc.), physiological (those of the bodies, e.g., breathing, coughing, etc.), and paralinguistic (those vocalizations which do not amount to intelligible language, e.g., mumbling, grunting, babbling, etc.). Interactive behavioural processes can be divided into perceptive (those performed by means of our senses, e.g., watching, staring,
sniffing, touching, etc.) and reciprocal (those which involve reciprocal movements of both participants, e.g., dancing, fighting, hugging, kissing, etc.). Behavioural processes are the least distinctive of all the process types. For example, behavioural processes are similar to material processes in that they require energy expenditure. However, they can be distinguished from material ones because their main participants are conscious.
1.2.3 Application of SFL
There have been numerous studies of different discourse types which used the framework of Hallidayan Systemic Functional Linguistics as a tool for various studies (Ansary & Babaii, 2005; Ji & Shen, 2004; Kyrala, 2010; Lins de Almeida, 2008; Norgaard, 2003; Oteiza, 2003;
Rodríguez, 2007). For instance, SFL approach was widely used in studies of discourse types such as literary discourse (Ji & Shen, 2004; Norgaard, 2003), media discourse (Ansary &
Babaii, 2005; Lins de Almeida, 2008; Rodríguez, 2007), political discourse (Kyrala, 2010), pedagogical discourse (Oteiza, 2003), and others.
For example, Ji & Shen (2004) in their study of Sheila Watson's novella "The Double Hook", used the model of transitivity as a useful tool for the interpretation of the text and the investigation of a protagonist's mental transformation. They analyzed the clauses of the text in terms of process types, their transitivity features, and the role of the novella's main character. As a result of the analysis, the authors suggested that there was a growing participation of the novella's protagonist in material, mental, and verbal processes. One of the studies of media discourse by Rodríguez (2008) used Halliday's (1994) SFL framework to explore functional variation across different situation types in comment articles from two British national newspapers, "The Guardian" and "The Sun". The author analyzed the relation of variations of language use and variation of social context in terms of Field, Tenor, and Mode. In the field of political discourse, the study by Kyrala (2010) analyzed the samples from speeches by John McCane and Barack Obama using the principles of SFL by Halliday
(2004). The author identified ideologies as well as differences and similarities of these two politicians by revealing and analyzing the use of experiential, interpersonal, and textual metafunctions in the texts which showed their perspectives on the economy and politics.
From these examples it can be concluded that SFL approach is widely applied to different discourse types in order to understand the relation of language use and social context. However, the application of SFL is not limited by the analysis of the discourse of non-brain-damaged speakers. There are numerous studies of pathological discourse with the method used within SFL framework (Mortensen, 1992; Sherratt, 2007; Armstrong, 2001, 2005).
For instance, Mortensen (1992) applied a transitivity analysis to discourse of a person with dementia of the Alzheimer's type. The study examined the clause structure within a discourse of the dementia client from a systemic-functional perspective. The focus of the study was on one aspect of SFL – experiential metafunction, described by Halliday (1994) as a resource to interpret and express our experience of inner and outer world. The experiential metafunction is expressed through naming which is reported to be impaired in the discourse of subjects with dementia of the Alzheimer's type. Mortensen used the systemic-functional model which was enable to analyze and synthesize the various aspects of language. In order to reach this goal, discourse samples of the dementia client were collected over two contexts:
procedural (description of how something is done) and recount (recalling of events and experiences). The author further analyzed three components of the resultant clauses: the use of processes, participants and circumstantial elements. Analysis of processes revealed the use of a wide range of process types with material being the most often used, followed by relational, mental, verbal, and behavioural. Analysis of participants in both contexts revealed infrequent omission of nominal groups. A similar picture was revealed about the circumstantial element by means of nominal groups' analysis within the prepositional phrase.
Thus, the results showed the impaired resource for the expressing the experiential meanings.
It was manifested in a wide range of all the elements use which were, however, simple, repetitive, and lack of specificity.
The analysis of another type of pathological discourse was done by Sherratt (2007).
Using SFL framework, she examinesd the ability of individuals with right brain damage (RBD) to verbally express their emotions. The study emphasized that language and emotions were related to each other and could not be easily separated into autonomous modules.
Verbal emotional expression (or evaluation) are particularly important in narratives because narratives without evaluation are flat and pointless. In RBD clients verbal emotional expression were hypothesized to be impaired because it was the right hemisphere that was dominant for emotional processing. In order to check this hypothesis, seven persons with RBD and ten non-brain-damaged speakers were recruited. Then two narratives (with positive and negative emotional discourse) were elicited and appraisal resources were further identified and classified by the author in terms of the appraisal categories (such as affect, judgment, appreciation, amplification, and so on). The results showed that a group of right hemisphere damaged speakers used less appraisal resources compared to their non-brain- damaged speakers. Additionally, the quantity and choice of verbal emotional expressions were impaired.
One more type of a pathological discourse was analyzed in a study of four individuals with aphasia by Armstrong (2005). The study described lexical patterns found in aphasic speakers' everyday discourse (such as recount) and examined their ability to participate in it.
In literature, increasing attention is given to verb usage that is reported to be impaired in speakers with aphasia. In the current paper, discourse meanings through the verb's role were examined from a systemic-functional perspective. Particularly, verbs were analyzed from a functional grammatical perspective as processes which realized them. Processes, by turn,
characterized different genres. Specifically, the recount genre was postulated to be characterized by predominance of material processes. Thus, the aim of the study was to find out the contribution of verbs to the discourse meanings produced by aphasic speakers. In order to reach this goal, four aphasic speakers and four matched controls were obtained. Then discourse samples in the recount genre were elicited and resultant verbs were further categorized into one of five categories of process types (material, relational, mental, verbal, and behavioural). Discussing the results, Armstrong discovered different patterns of the process use across two groups of normal and aphasic speakers. Namely, while normal speakers used either predominantly material verbs, or material and relational verbs about equally, two of four aphasic speakers used more material verbs than relational and mental.
Armstrong explained that aphasic speakers were disable to include personal opinion or evaluation into their recounts. In addition, normal speakers demonstrated greater variety of verbs than persons with aphasia. Thus, compared to normal speakers, aphasic speakers produced the different patterns of verb usage in terms of verb types and their variety. This information suggested that persons with aphasia had a deficit in conveying meanings in discourse.
1.2.4 Challenges of applying SFL to Russian
The grammatical systems of Russian and English are fundamentally different. English is an analytical language, in which grammatical meaning is largely expressed through the use of additional words and by changes in word order. Russian, on the other hand, is a synthetic language, in which the majority of grammatical forms are created through changes in the structure of words, by means of a developed system of prefixes, suffixes and inflectional endings which indicate declension, conjugation, person, number, gender and tense. Russian therefore has fairly complicated systems of noun and adjective declension and verb conjugation, which allows the Russian sentence to have flexible word order.
The Russian and English verb systems express rather different kinds of meaning.
Russian has only three tenses: past, present and future. The verb system is mainly built on the notion of aspect. This is the contrast between actions which are incomplete (imperfective aspect) and those which are complete (perfective aspect). These contrasts are indicated through affixation. Perfect and progressive forms of verbs, as understood in English, do not exist. Strictly speaking there are no auxiliary verbs like do, have or will in Russian. Thus, the verb systems of Russian and English are different as well. This fact is of importance to the current study since the discourse analysis which is applied to it, has to do with verbs.
In SFL, Halliday (2004) describes five process types (material, relational, mental, verbal, behavioural) which are realized through verbs in clauses. In spite of the differences between English and Russian verbal systems, Russian translations of English verbs (and the opposite: English translations of Russian verbs) represent the same process types. For example, the English verb go is the material process type as well as the Russian translation idti. Similarly, the Russian verb dumat’ is of the mental process type just like its English translation think.
However, there is one difficulty of applying SFL system of processes to the Russian language: the verbs be (byt’) and have (imet’), which realize the relational process type, have more complex representation in Russian than in English. According to Tuniks (1972), byt’ is the Russian infinitive of the verb be. In the present tense there are only two forms: est’ which is the third person singular, and sut’ which is the third person plural and is more a different stylistic form of est’ remaining from Old Russian. For example, the verb be in a sentence "A straight line is the shortest distance between two points" in Russian is realized as "Pryamaya liniya est’ kratchaishee rastoyanie mezhdu dvumya tochkami". An example of a very rare Russian form of plural verb be can be represented in the following sentence: "Samye obyknovennye obshie otvlecheniya sut’: svoboda, ravenstvo, …, kul’tura" ("The most
ordinary common diversions are freedom, equality, …, culture"). In the future tense there is a complete paradigm of the to be forms: budu (I am), budesh (you are), budet (he/she/it is), budem (we are), budete (you are), budut (they are). In the past tense the forms differ only according to number and gender: byl (masculine form of the verb to be, singular), byla (feminine form of the verb to be, singular), bylo (neutral form of the verb to be, singular), and byli in the plural. The imperative forms are bud’ in the singular or familiar, and bud’te in the plural or polite.
Chvany (1970) suggests three main types of the Russian verb to be (byt’) according to their functions: byt’ of existence, byt’ of occurance, and byt’ of location. Byt’ of occurrence is non-stative, while the others are stative. Stativeness is determined not only by contextual features, but also by the mandatory absence of byt’ where it is stative. For example, "Ivan ***
doma" ("Ivan is at home"). Such a phenomenon of the obligatory absence of the Russian verb to be (byt’) is widely distributed and used, especially in everyday conversations.
Thus, in the present study, the application of SFL system of process types to Russian and categorization of verbs which realize the processes will be of special concern as it can influence the results of the research. It should also be noted that this study represents the first time that a transitivity analysis will be applied to the Russian language.
1.3 Genres of recount and opinion
People use a variety of genres in their daily communications in order to realize certain communicative goals. The linguistic definition of genre is provided by Russian theorist Bakhtin (1986) in which genres are identified as “relatively stable types of interactive utterances”. According to Martin (1997), genre is a staged activity in which speakers are members of their culture and have a certain purpose. Thus, genre is a cultural concept resulting from context. In addition, genre has a certain structure which identifies it.
Hyland (2002) suggests two assumptions to base on for identifying genre. First, texts belong to the same genre if their features depend on the social context of their creation and use. Second, those features should be described in a way that one text can be related to other texts like this in order to ascribe them to the same genre. Swales (1990) also provides a useful model for defining genres in terms of communicative purpose of discourse. According to his definition of genre, it represents a class of communicative events in which their members share some communicative purposes. These purposes provide the rationale for the genre. This rationale, in its turn, forms a schematic structure of discourse and determines the choice of its content and style. Thus, a communicative purpose are both criteria and the basis for rhetorical action to be focused on within the same genre. However, a communicative purpose is not the only criterion which defines genre. Other features which identify genre are similar patterns of structure, style, content, and intended audience. If all the criteria are realized, then the exemplar is considered to be prototypical in terms of a genre.
In SFL, genre is identified as a way of using language to achieve a certain culturally established goal. Therefore, genre is related to the context of culture (extralinguistic stratum of context). The structure of genre is a macro-view of syntagmatic aspect. Halliday (1985, 1994) provides such a model of genre in which the contextual dimensions of genre are related to the semantic and grammatical organization of language itself. Additionally, SFL has the potential to develop features and structures of different genres and provides criteria for how genres can relate to each other. Thus, different genres in a single culture and similar genres across cultures can be accounted for the SFL model.
Different linguists distinguish different genres. For example, Martin & Rose (2003) define the following typical genres: service encounters, casual conversations, arguments, telephone inquiries, games, jokes. Another examples of genres provided by linguists can be
procedure, anecdote, description, exposition, recount, report, discussion (Gerot & Wignell, 1994). In the current study, two genres are chosen: personal recount and opinion.
Genre of personal recount
Gerot & Wignell (1994) define a personal recount as a genre in which events are retold for the purpose of informing or entertaining. An event or an experience described in a personal recount takes place in the past. Rothery (1994) also notes that a recount can be some unusual or problematic event which is not resolved in the end (unlike narratives). Examples of personal recounts can be personal stories of a wedding event, a birthday party, or a reunion.
A personal recount consists of three main elements: social function, generic structure, and significant lexicogrammatical features. All of these elements should be present in a genre for it to be considered a personal recount. The first element is social function which is a communicative purpose of speakers, namely, information or entertainment. The second element of a personal recount is generic structure. Hasan (1985), for example, suggests a structure with several stages of genre of a personal recount: orientation, record of events, reorientation, and coda. At the stage of orientation, speakers' names, ages, appearances, purposes of communication as well as description of weather and place may be involved. The stage of record of events consists of speakers' emotions, thoughts and feelings about a particular event. The reorientation stage involves the summary of a recount, i.e., the ending of a story. At the stage of coda, speakers give their attitude, evaluation of events. The third element of a personal recount is lexicogrammatical features which are certain participants, circumstances of time and place, use of past tense, and temporal sequence.
Genre of opinion
According to van Dijk (1982), the genre of opinion implies that language users in natural communicative situations use not only their knowledge about persons, objects, events, actions, but also use their evaluative beliefs. Opinion genre also involves evaluations and is related to values and norms. They are notions that are traditionally linked to reputedly vague things like affect and emotions. Thus, representation of the genre is subjective. Shi-xu (2000) gives a definition of the opinion genre as a kind of genre in which individuals or groups present a mental, subjective, and personal belief. Belief can be about anything, real or imaginary, but is often about a state of affairs (how things are) or a course of action (what should be done). Furthermore, this personal belief can be influenced and interconnected with different kinds of the social ones. Greenberg (2000) says that the opinion genre is not intended to be objective, fair, or balanced. In contrast, it is consisted of overtly biased viewpoints.
Shi-xu (2000) describes three main characteristics of the opinion genre. The main characteristic of the opinion genre is a sense of being subjective, as opposed to objective. In this frame, distance from reality is often suggested. Another feature of the opinion genre is that it is personal and individual. It means that opinion is offered as one's own, separate and different from others' opinions and usually does not try to influence them. A final characteristic of the opinion genre, according to Shi-xu, is a sense of it being a mental trait. It implies that speakers offer their opinion to suggest that their opinions are what they think or represent something in their minds. One more feature, proposed by Somasundaran, Wiebe, &
Ruppenhofer (2008), is the polarity of the opinion genre, which means that opinions can be positive and negative. Greenberg (2000) also adds such a characteristic of the opinion genre
as its evaluative nature, which means that in giving their opinions, speakers take sides by evaluating and explaining events.
Somasundaran and colleagues (2008) suggest two types of opinion genre: sentiment and arguing. Sentiment includes positive and negative evaluations, emotions, and judgements. Arguing includes arguing for and against something, and arguing that something should or should not be done. Another division of opinion genre is proposed by van Dijk (1982). He distinguishes personal and group (or public) opinion where individual opinion is one's assignments of values to objects, persons, states, events, and actions, whereas group opinion implies that people share the same opinion about some objects, states, events, and actions. Additionally, van Dijk distinguishes between personal opinion and social opinion.
The social type of opinion genre is inferred from general norms or values.
1.4 Hypothesis and predictions
1.4.1 Process production variation within genres
Different types of processes have been shown to be used more often or less often depending on the genre (Halliday, 1994; Mortensen, 1992). Research in text linguistics suggests that different genres of text, such as recount and opinion, may be characterized by different process types (Armstrong, 2001, 2005; Eggins & Martin, 1997; Halliday, 1985).
In the genre of recount, it has been proposed that it is characterized by the dominance of the material process type (processes of "doing" and "happening", such as run, bake). In addition, another frequently used process type is relational (processes of "being" and
"existing", realized by the verbs be, have). For example, Armstrong (2001) conducted a study of process production variation in the genre of recount by eliciting narratives on the topics of illness, wartime experience, job history, and happy event. The results revealed that the normal speakers used either predominantly material processes, or they used material and
relational process types about equally. Their texts were recounts of events through the use of material processes with some elements of opinion and evaluation (they were realized by relational processes).
Thus, the genre of recount, being a description of an event or an experience which take place in the past, imploys the most frequent usage of the material process type. However, this genre, being personal recount, also contains elements of opinion and evaluation, and they are realized through the frequently used relational processes. Concerning the genre of opinion, it has been proposed that it is characterized by the dominance of relational and mental processes (processes of sensing, such as think, feel). For instance, in a study by Armstrong (2005), the ability of aphasic speakers to express feelings, opinions, and attitudes was examined. Five non-brain-damaged speakers were asked to talk about a happy event, an illness they had experienced, and what their jobs had been involved. The results showed that the most frequently used process types were mental and relational processes.
Thus, defined as mental, subjective, and personal belief, the genre of opinion implies the most frequent usage of mental and relational processes as they realise sensing, cognition, and perception of a state of affairs or a course of action. However, the material process type is supposed to be frequent as well, since narratives in the opinion genre often contain elements of description which are realized by material processes.
Based on the assumptions given above, process production variation within two genres is hypothesized to be the following: (1) the genre of recount would retrieve the highest frequency of material and mental processes; (2) the genre of opinion would retrieve the highest frequency of relational, mental, and material process types.
1.4.2 Process production variation across genres
It is assumed, that the genre of recount requires recounting and description of actions, events and happenings. It is realized by material and relational processes. At the same time, the
genre of opinion requires expressing evaluation and opinion, i.e., conveying feelings. These are characteristic of mental and relational processes.
Based on these assumptions, it can be expected that across two genres of recount and opinion the pattern of process production variation would be the following: (1) material processes would be more frequently used in the genre of recount than in the genre of opinion;
(2) relational processes would be of higher usage in the genre of opinion than in the recount genre; (3) mental processes would be more frequent in the opinion genre than in the genre of recount. Two other process types (verbal and behavioural) would be the least frequently used in both genres.
Chapter 2 Method
In order to test the hypothesis of the current study and explore differences and similarities in the use of different process types across discourse genres, ten native Russian speakers were recruited, and were asked to discuss two different topics constituting recount and opinion genres. Transitivity analysis based on Systemic Functional Linguistic approach proposed by Halliday and Matthiessen (2004) was used to analyse the resultant discourse. The choice of transitivity analysis within the SFL framework was not random. While there are different frameworks which help to examine clause level meanings, transitivity analysis captures different semantic elements of the clause and relates them to overall discourse patterns. Thus, transitivity analysis is the grammatical resource for how people use language in their daily life. Spontaneous speech in a form of monologues was elicited in two different discourse genres (personal recount and opinion) to derive different process types for further analysis.
2.1 Participants
Participants were recruited in the South-Western part of Siberia, Russia. There were five female and five male native speakers of Russian all of whom met the inclusion criteria.
Inclusion criteria were that participants were monolingual non-brain-damaged Russian speakers who had no history of neurological or psychiatric disorders. The ages of the participants ranged from 49 to 56 years (M = 51.8). Such an age range of non-brain-damaged participants corresponds to a normal age range of aphasic speakers. For example, in a study reported by Armstrong (2005), the age of aphasic participants ranged from 51 to 69 years (M
= 61.8). All the participants of the current study had 15 years of education (see Table 2.1).
Table 2.1. Participant characteristics.
Subject Sex Age Year education Occupation
N1 N2 N3 N4 N5 N6 N7 N8 N9
N10
F F F F F M M M M
M
56 51 54 52 49 52 52 51 51
50
15 15 15 15 15 15 15 15 15
15
teacher accountant accountant engineer businesswoman
foreman business owner company lawyer
director of the charitable nonprofit
foundation teacher
2.2 Discourse elicitation procedure
In order to elicit discourse samples in two different genres (personal recount and opinion), participants were asked to give monologues resulting in two different samples per participant.
First, participants were asked to give a five-minute speech sample recounting a happy event
they had experienced and the circumstances surrounding it. The following question was used to elicit this discourse sample: "Please, tell me about a memorable happy event of your life.
Support your story with details of that event". Occasional encouragement by the interviewer was signalled by "mmm", "ok". However, if a speaker's story was too short, further prompting was used. The interviewer then asked: "Could you add some more details to your story?" Although three participants initially gave a very brief recount of a happy event they had experienced (no more than two minutes in length) and further prompting was used, the majority of the participants did not need any prompting and gave sufficient discourse samples (about five minutes each). After recounts were elicited, the participants were then asked to give their opinion on the topic "Silicon plant construction on the territory of Abakan". The choice of such a topic is due to its topicality and direct relationship to the participants. This topic requires participants to express their opinion about it. The following question was used to elicit this discourse sample: "You know that the silicon plant construction is planed in our city. Please, express your opinion on this issue and explain it". The majority of the participants gave sufficiently long samples in this context. However, four participants required further prompting by asking "Could you give other arguments in support of your point of view".
All monologues obtained at the participants' homes resulted in the collection of a total of 20 story samples, 10 in each genre: recount and opinion. Each speech sample was about 5 minutes long and 63 clauses in length. All discourse samples obtained were audio-taped and later orthographically transcribed.
2.3 Data transcription
The language samples were divided into clauses which are the basic grammatical units in the Systemic Functional Linguistic (SFL) approach by Halliday and Matthiessen (2004). The
meanings the clauses convey depend on which elements or functional components of the clause (i.e., participant, process, and circumstance) are chosen by the speaker. The processes are the obligatory components of all clauses in the transitivity system and they express meanings related to doing, happening, saying, thinking and being (Halliday & Matthiessen, 2004). Processes are realized by main verbs, for example, ‘go’ or the verbal group, for example, "have to get going". Participants are typically realized by nominal groups.
Circumstances describe aspects of location, time and manner which are usually realized by adverbial or prepositional groups. The details about process types are described in the Introduction Chapter.
The division of a text into clauses started with identification of a process. Each simple clause has one process and can stand alone, that is, it is an independent clause. For example, a sentence "If you stumble or do something else there, they will shoot" (Participant N1;
discourse genre of personal recount) has two identifiable processes (stumble and shoot), and hence was counted as two independent clauses and transcribed on separate lines, with the processes underlined:
-1- Esli tam Ø spotkneshsya ili chto
If there (you) stumble or something
-2- Ø Budut strelyat' (they) will shoot
"If you stumble or do something else there, they will shoot"
During the process of clause division there were a number of additional considerations. Some of these were mazes. Mazes were filled pauses, false starts, repetitions, reformulations, interjections, and parenthetical words and expressions. Certain mazes formed a part of the transcription and were not counted as individual clauses in the analysis. An example of a parenthetical maze produced by one of the participants (Participant N1;
discourse genre of personal recount) is illustrated in -3- (process is underlined, the parenthetical expression is in bold):
-3- Nu i poslednim etapom tak skajem zaversheniem etogo kruiza bila poezdka Well and last stage so to say completion this cruise was trip
"Well, and the last stage, so to say, the completion of this cruise was a trip"
In this example, the process tak skajem ("so to say") was used as a parenthetical expression in order to specify the topic of the current sentence. Although this process can be identified as verbal process, it was not counted in the study because it did not contribute to the message.
Another example of a maze which was not counted during the analysis was repetition as illustrated in -4- (Participant N7; genre of personal recount):
-4- Nu Ø brali [brali] palatku s soboi na vsyakij sluchaj Well (we) take [take] tent with us just in case
"Well, we took a tent with us, just in case"
Here, the participant used the process brali ("take") twice, however, it was counted only once as the participant used it unintentionally, as the pause filler trying to recall the details of his story.
One more example of a maze which was not counted as a clause during the analysis was when the participant tried to involve the interviewer in the monologue using imperative or interrogative sentences. For example, one of the participants (Participant N5; discourse genre of recount) while telling the story of her daughter's wedding, used an imperative sentence addressing the interviewer:
-5- Da vot predstav'!
Yes just imagine!
"Yes, just imagine!"
Although it is possible to have a clause realized by the process predstav' (‘imagine’), in this case the participant tried to share the emotions by engaging the interviewer, and this expression could not be counted as a part of a personal recount.
Another consideration which was taken into account during the process of clause division and a transitivity analysis was the frequent usage of zero copula. As it was already mentioned in the Introduction Chapter, there is a tendency in the Russian language to omit the copula byt' ("be") in the present tense. Thus, for data transcription the elliptical form of the relational process "be" was counted since it has a function in a clause even without its form. An example of such a phenomenon is a sentence used by one of the participants (Participant N6; discourse genre of opinion) where the elliptical form of the relational process presented by the verb "be" is indicated as three asterisks:
-6- Nashe otnoshenie konechno *** otrizatel'noe Our attitude of course is negative "Our attitude is, of course, negative"
In this sentence, although the verb "be" has an elliptical form, it was counted as a relational process because it contributes to the main meaning of the message.
An additional minor consideration not essential for process type identification but important for clause division was detection and notation of a so-called "null subject" which is also a characteristic feature of the Russian language. The null subject was the Participant component of a clause and therefore was marked in order to identify the boundaries between clauses, as illustrated in -7- (Participant N1; genre of opinion):
-7- I Ø znayu rezul'taty And (I) know results
"And I know the results"
Even though the process type is identifiable in any case, the marking of the null subject helped to make the clause independent and clear.
2.4 Data analysis
After orthographical transcription of the discourse samples and division into clauses was completed, the samples were then analysed by process types according to Transitivity Analysis (Halliday & Matthiessen, 2004).
A transitivity analysis (Halliday & Matthiessen, 2004) was completed as a measure of process variability. The procedure for categorization of process types started with identifying the process, i.e., the main verb or verbal group in each clause. Each simple clause had one process which was counted in the analysis. Then the process was categorized into one of five main groups: material, relational, mental, verbal, or behavioural. Once all processes used by the speakers were categorized, then these were tallied to provide a total number of processes for both genres (recount and opinion) and a total of each type of processes for each of the speaker’s language samples. Every type of process was then calculated as a proportion of total processes used in the genres of personal recount and opinion. An example of speech sample analysis derived in the current study is given in Appendix 2.
Finally, inter-rater reliability was checked using a random sample of 20% of language samples for the transitivity analysis by a second researcher. Percentage inter-rater agreement for the procedure of total categorization of processes was counted. Inter-rater disagreements were discussed and a consensus decision was reached where possible. One consideration taken into account during categorization of processes was in rare cases of discrepancy in process types of Russian verbs and their English translations. In such cases, the discussion with native English-speaking and native Russian-speaking experts took place, and a consensus decision was reached.
Chapter 3 Results
3.1 Sample length
Obtained speech samples ranged in length from 45-88 clauses (M = 63). The average number of clauses produced by the speakers during the personal recount was 70.1 clauses (range 51- 88) while the average number of clauses the participants produced during opinion was 55.9 clauses (range 45-77). The performance of female participants resulted in an average of 76.6 (range 69-88) clauses in the genre of recount and 56.8 (range 48-74) clauses in the genre of opinion. The length of obtained speech samples of male participants was 63.6 (range 51-73) clauses in the genre of recount and 55 (range 45-63) clauses in the genre of opinion.
Individual participant's data for each genre are presented in Table 3.1.
3.2 Statistics
Two statistical methods were used to answer the study questions. Two main goals were set in the current study of unimpaired population. The first goal was to investigate process production variation within genres. The second goal was to investigate process production
Table 3.1. Number of clauses produced by each participant in each genre.
Participant Recount Opinion
F1 F2 F3 F4 F5 M6 M7 M8 M9 M10
69 88 77 74 75 51 73 70 72 52
48 74 55 56 51 45 63 58 48 61
variation across genres. The variation of process production within genres was assessed using a one-way ANOVA which tested the significance of scores' percentage across process types within each genre. The variation of process production across genres was assessed using a paired-sample t-test which tested the significance of scores' percentage for each process type across two genres.
3.3 The pattern of process use in the genre of recount
In the genre of recount the participants produced 701 processes overall, 389 (55%) of which were material, 181 (26%) were relational, 70 (10%) were mental, 22 (3%) were verbal, and 39 (6%) were behavioural processes (see Figure 3.1).
Material, relational, and mental types were the most frequently used processes compared to verbal and behavioural ones in personal recounts. This pattern of process use
Figure 3.1. Percentage of process types in the genre of recount.
was found consistently across all the participants. In general, speakers used more material processes than all the other four process types combined. The verbal type was the least frequently used process.
A one-way ANOVA analysis was performed on the data in order to further investigate whether the differences across process types were statistically significant. The results indicated that the differences in frequency of process types used in the genre of recount were significant, F(4,45) = 98.157, p < .01. A post hoc test revealed that the difference across material (M = 55.9, SD = 3.6), relational (M = 25.9, SD = 2.6) and mental (M = 9.8, SD = 1.2) process types was significant at p < .05. Verbal (M = 3, SD = 0.8) and behavioural (M = 5.4, SD = 1.6) process types did not differ significantly from each other and from the mental process type. Material and relational processes differed significantly from verbal and
material relational mental verbal behavioural
0 5 10 15 20 25 30 35 40 45 50 55 60
behavioural process types at p < .05. These results reflect the fact that because the genre of personal recount is a descriptive listing of events, it presupposes the highest usage of material processes of "doing" and "happening". In addition, relational and mental processes were also used quite frequently. Thus, the genre of personal recount contains not only a component of events listing but also a component of personal evaluation which is characteristic of relational and mental process types.
An example of the pattern of process use in the genre of recount can be illustrated by a text produced by one of the participants (Participant F4). In the text below, Participant F4 most frequently used material, relational, and mental process types as she was telling about a trip she had (process types are in bold, processes are underlined).
(material) S'ezdili my na "Sunduki"
Go (1st,pl,pres) we to "Sunduki"
"We went to the “Sunduki”"
(mental) Poezdka Ø ponravilas' ochen'
Trip (I) like (1rd,sg,past) a lot
"I liked the trip a lot"
(relational) Emotzij bylo ochen' mnogo
Emotions w as (3rd,sg,past) a lot
"There was a lot of emotions"
(material) Mashinu my svoyu ostavili doma
Car we our (3rd,pl,pos) leave (1st,pl,past) at home
"We left our car at home"
(material) Ø Poehali na “GASeli”
(We) go (1st,pl,past) by “GASelle”
"We went by “GASelle”"
Other speakers in this study demonstrated a similar pattern of process use during their personal recounts.
3.4 The pattern of process use in the genre of opinion
In the genre of opinion the participants produced 559 processes in total. Among them 246 (44%) processes were material, 184 (33%) were relational, 83 (15%) were mental, 33 (6%) were verbal, and 13 (2%) were behavioural processes (see Figure 3.2).
Figure 3.2. Percentage of process types in the genre of opinion.
The most frequently used process types in the genre of opinion were material, relational, and mental, whereas verbal and behavioural processes were of least frequency. A one-way ANOVA analysis showed the significant differences across process types used in the genre of opinion, F(4,45) = 70.922, p < .01. A post hoc test revealed that only the difference between verbal (M = 5.7, SD = 1.2) and behavioural (M = 2.4, SD = 1.1) process types was not significant at p < .05. Material (M = 44.2, SD = 2.5), relational (M = 32.7, SD =
material relational mental verbal behavioural
0 5 10 15 20 25 30 35 40 45 50
3.2) and mental (M = 15, SD = 1.9) process types differed significantly from each other as well as from verbal and behavioural processes at p < .05.
The prediction of the current study was in favour of the highest frequency of relational processes in the genre of opinion that is defined as expression of a point of view, evaluation. It was found, however, that the participants had a stronger preference for material over relational processes. One possible explanation for these results could be the influence of the topic "the silicon plant construction on the territory of Abakan". This topic seemed to be appropriate for the genre of opinion since it met the main characteristic of the genre and had direct relevance to the participants. Thus, the participants were supposed to express their attitude to this topic and give reasons for it and it was hypothesized that they would use significant amount of relational and mental processes to achieve this. Indeed, the speakers expressed their opinion and gave arguments to support it. However, many participants used not only reasoning and argument, but also personal recount as a way to support their point of view. For instance, Participant M6 supported his argument against silicon plant construction by presenting his colleague's experience of working near a silicon plant. Below a text of Participant M6's speech is presented (with process types in bold and processes underlined) where the most frequent process type is material.
(material) Na rabote chelovek rabotal vozle kremievogo zavoda
At work person work
(3rd,sg,past)
near silicon plant
"At work one person worked near the silicon plant"
(material) Nu on uzhe stalkivalsya
Well he already encounter
(3rd,sg,past)
"Well, he already encountered with this"
(material) Uzhe Ø rabotal na kremnievom zavo
de
