Showing papers in "Text - Interdisciplinary Journal for the Study of Discourse in 1989"

Styles of stance in English: Lexical and grammatical marking of evidentiality and affect

Abstract: The present paper identifies anddescribes various speech styles of English äs marked by stance. By stance we mean the lexical and grammatical expression of attitudes, feelings, judgments, or commitment concerning the propositional content of a message. In an earlier paper (Biber and Finegan, 1988), we limited our investigation to the adverbial marking of stance; here we extend the analysis to include adjectival, verbal, and modal markers of stance. All occurrences of a large sei of stance markers are identified in 500 texts, drawn principally from the LOB and London-Lund corpora (of written and spoken British English). The stance markers are divided into 12 categories basedon semantic and grammatical criteria, and thefrequency of occurrence for each category in each text is computed. The twelve categories are (1) affect markers (adverbs, verbs, and adjectives); (2) certainty adverbs; (3) certainty verbs; (4) certainty adjectives; (5) doubt adverbs; (6) doubt verbs; (7) doubt adjectives; (8) hedges; (9) emphatics; (10) possibility modals; (11) necessity modals; and (12) predictive modals. Using a statistical technique called düster analysis, texts that are maximally similar in their exploitation of stance markers are sorted into clusters. We Interpret each düster äs a stance style by consideration of the predominant stance jeatwes in the düster, the situational characteristics of the texts constituting the düster, and a functional analysis of individual texts. Overall, six stance styles are identified, among which are 'Emphatic Expression of Affect', 'Expository Expression of Doubt', and 'Faceless'.

Language has a heart

Abstract: In the past several years, the social sciences have been articulating how emotion impacts cognition and social action. Linguists have underestimated the extent to which grammatical and discourse structures serve affective ends. A cross-linguistic analysis indicates that languages dedicate phonological, morpho-syntactic and discourse features to intensify and specify attitudes, moods, feelings and dispositions. These features provide an affective frame for propositions encoded. Suchframes can be consideredas pari of the Information expressed, äs affective comments on the expressed propositions they address. These comments Interface with gestural cues to provide interlocutors with critical Information on which to base subsequent social actions.

Literacy and feelings: The encoding of affect in Nukulaelae letters

Abstract: The primäry purpose of literacy production on Nukulaelae Atoll (Central Pacific) is to write letters. Nukulaelae letters are sent to relatives on neighboring atolls, and serve a variety offunctions: monitoring economic reciprocity; informing kin of family events; and admonishing younger people. Permeating every aspect of letters is a heavy emphasis on the overt expression of affect, of a nature notfound in any other arena of Nukulaelae social life. This paper describes the way in which affect is encoded in the text of these letters, and shows how the topics addressed in letters are emically defined äs affectively charged. A content and historical-ethnographic analysis of letter-writing on the atoll indicates that letters have been defined äs cathartic events from the very introduction of literacy. It is suggested that the metaphorical affiliation of letter-writing with parting is in large pari responsihle for letters having become qffect-display context. This case study challenged traditional views that Dritten communication is universally less affective than spoken communication. Tapauta, Sept. 13, 1897 To Mrs. David, the lady,— My love to you! alas my mother! The thought weeps when I think of you, together with the others, because of your kindness to me. Alas for my love! Dear, oh dear, my heart is füll of love, but it is difficult because I cannot speak; but I thought I would try and send this small piece of paper to make known to you my love. Alas my mother! my love is very great, and it is difficult and hard because we shall be so soon parted. Grief continues to grow in my heart when I think of the days we were together in Funafuti. Alas! I do not forget them and you all. I feel I want to be still with you. It is hard that we have been so soon parted on shore. May you return with blessing to your home. This love of mine has nothing with which to make itself known, but I have striven to make appear before you that which was hidden, namely, my love to you. Alas, my parents, love is difficult. 01655-4888/89/0009-0069 $2.00 Text 9 (l) (1989), pp. 69-91 © Mouton de Gruyter

Telling age in later life: Identity and face implications

Abstract: The disclosure of age-in-years in later life is a frequent characteristic of intergenerational talk among first-acquaintances. Qualitative analyses of particular sequences show that considerations of age and of appraised health/wellbeing are structurally interrelated in these disclosures. Through this relationship, chronological age can function as a token in the bilateral negotiation of age-identities in discourse. Despite its ritualised character, age-disclosure can impact significantly on elderly people's self-esteem and on cross-generation attitudes. Variably, at different points in the life-span, taboos and normative prescriptions are associated with both seeking and providing information about age. Personal experience (in the absence of empirical studies) suggests that children and adolescents are often asked to tell their age by distant relatives or new acquaintances, or have this information revealed for them by parents or guardians. Inquiring about age in these contexts, to the extent that it is more than ritualised, seems to be part of a signalling of engagement and perhaps nurturance by well-meaning adults interested in following the growth and development of the young. In the middle years of life, age-in-years, and perhaps all discussion of own age, drops out of unmarked usage, appearing to feature predominantly in mockdenigrating remarks about the passing of time, references to the unwanted arrival of birthdays, commercially promoted through the greetings-card industry. Age-talk finds its place as part of a broad spread of ageist constructions in Western society across several domains — literature (Berman and Sobkowska-Ashcroft, 1986), humour (Palmore, 1971), magazine fiction (Martel, 1968), television drama and commercials (Kubey, 1980). The admixture of fear, reticence and regret with which, facetiously or 01655-4888/89/0009-0129 $2.00 Text 9 (2) (1989), pp. 129-151 © Mouton de Gruyter 130 Nikolas Coupland, Justine Coupland and Howard Giles not, many middle-aged adults appear to represent their own ageing, and the consequent teasing and chiding of those whose ageing comes up for review, undoubtedly form part of the interactional means by which negative images of ageing and the elderly are reproduced (cf. Kenwood, N. Coupland, J. Coupland and Giles, submitted). Images crassly invoked here amount to a checklist of negative elderly stereotypes: frailty, sexual inactivity, incompetence, grouchiness, unsociability, and so on (Braithwaite, 1986; Giles, N. Coupland, Kenwood, Harriman and J. Coupland (in press); Stewart and Ryan, 1982). Within this calendar-marking tradition and associated discourses, personal life-spans are viewed as incremented scales with natural boundaries which are bench-marks to decrement: physical, social and socio-psychological (cf. Branco and Williamson, 1982; N. Coupland and J. Coupland, in press b). Decade-boundaries seem to have a particular salience in this respect, as do transitions from one generational category to another (parent to grandparent, for example). On the other hand, in old age, the life-span territory we explore in this paper, age-in-years re-surfaces from its underground life. The data we draw on, from two audio-recorded contexts, confirm the prevalence of disclosing chronological age (DCA) among at least some elderly groups. Given the pejorative associations of advancing age among younger groups (as we have speculatively sketched them), we are led to ask why elderly DCA might be not only tolerated but positively construed on occasions. The particular aim of the paper is therefore to consider the social functioning of DCA in context, and to achieve this through examining the discursive sequences and routines through which, in our data, age-disclosure is managed. Our interpretations are couched in terms of age-identity, and more particularly the various elderly-identifications that DCA can encode. For this reason, we begin with a brief consideration of the diverse literatures that have been concerned with age-identity and age-marking in later life. Approaches to age-identity An early experimental psychological tradition (for example, Tuckman and Lorge, 1954; Riley and Foner, 1968) was concerned to assess elderly self-concepts, and found a common tendency for ageing people to 'deny' their elderliness, judging themselves as younger than their chronological age in years. More recent research has explored different interpretations of such age-categorizations (see, for example, Guptill, 1969; Mutran and Burke, 1979), though methodological criticisms have also been made (cf. Telling age in later life 131 Breytespraak and George, 1979). A recent, more global 'life-position indicator' is Rubin and Rubin's (1982; 1986) index of'contextual age', a questionnaire-based measure aggregating objective and subjective dimensions of elderly people's health, mobility, finances, social networks, and other variables. On the other hand, symbolic interactionists (see, for example, Spence, 1986) and Meadian theorists (Chappell and Orbach, 1986) have argued for studying the ageing self as process rather than in static terms. It is this tradition of concern for 'identity-in-the-world' (Ainley and Redfoot, 1982) which is most receptive to the discourse analytic approaches to identity that Potter and others (Potter 1988; Potter and Wetherell, 1987) have argued for. (For a more detailed review, see J. Coupland, N. Coupland, Giles and Kenwood, in press) Still, linguistic studies of age-categories and age-references are rare. There have been some literary and dictionary analyses (Berman and Sobkowska-Ashcroft, 1986; Covey, 1988), and occasional experimental studies (for example, Nuessel, 1982; Rodin and Langer, 1980). Barbato and Feezel (1987) asked people in three age-bands (17-44, 45-64, 65 + years) to rate a list of 10 terms referring to an elderly person on five semantic differential scales. Predictably, terms including 'senior citizen', 'retired person' and 'mature American' were more favourably perceived than, for example, 'elderly', Old', and Old folks'; there were few differences between the responses of different age-groups. Research of this sort needs to confront the criticisms (for instance of Potter and Wetherell, 1987, ch. 6) that group-attributes and even social categories themselves are not static givens, and can be constructed out of the socio-psychological needs and goals of interactants in discourse. As a result, the objective of obliterating ageist terms from the language (Oyer and Oyer, 1976) may be misguided, ignoring the essential creativity of group-attributions and hence ageism in context, and their consequent resistance to change (cf. Singh, Lele and Martohardjono, 1988; Snyder, 1981). In a recent paper (J. Coupland, N. Coupland, Giles and Henwood, in press), we have attempted to overview the range of processes by means of which age-identity is marked, variably and negotiatively, in intergenerational talk. We have identified two broad categories of such processes. First, 'Age-Categorisation Processes', subsuming direct and indirect references to the elderly participant's group-membership, time of life, generational role, etc., identification of frailty and age-associable problematical circumstances (cf. N. Coupland, J. Coupland, Giles, Henwood and Wiemann, 1988), and the disclosure of chronological age. As a second general set, we consider 'Temporal Framing Processes', which include the adding of a time-past perspective to current issues (cf. Boden and Bielby, 1986), elderly people self-associating with the past, and the 132 Nikolas Coupland, Justine Coupland and Howard Giles recognising of historical/cultural/social change. Intergenerational talk in the first-acquaintance interactions under investigation was in fact imbued with interweaving age-identity-marking processes under these headings, though highly specific contextualisation possibilities and implications are associated with each dimension. Our earlier overview has been unable to explore the sequential characteristics of each of these processes in individual sequences of talk, though it is, we argue, at this level that interpretive inferences of goals and consequences are most secure. It is against this background that, in the present paper, we examine the management of chronological age-disclosure as a particular and particularly focused means of negotiating elderly identity in conversation.

‘Sure, sure’: Evidence and Affect

Abstract: When people fight, they typically display their feelings; and since people often fight over matters offact, declarations of truth and accusations of deceit are frequent vehides for affective language. I examine evidential devices in Tzotzil arguments, to display the complex formal and functional intermingling of referential usage (the normal accepted domain for questions of truth), rhetoric andgenre (in particular, the highly charged formal speech style in Zinacanteco Tzotzil), illocutionary force, and affect. I conclude that (Tzotzil) conversation äs much establishes a moral äs a propositional universe of discourse, and that evidentials can be about feelings and commitments äs well äs about truth. Fussing and fighting On the radio, the Bickersons are at each other again: John: Blanche, what's the matter with you it's three o'clock in the morning You had a good time t'night Now whyncha let me sleep? Blanche: had a miserable time it was the UNhappiest anniversary I ever spent Why didn't you show up for the party, John? = John: = I TOLD ya I got stuck at the office Blanche: Fd like to believe that What were you doing? John: working Blanche: su:re sure That's always the first excuse If I don't fall for that? 01655-4888/89/0009-0027 $2.00 Text 9 (l) (1989), pp. 27-68 © Mouton de Gruyter

Forms and footings in boys’ sportscasting

Abstract: This paper describes the way in which three school-age boys act äs sportscasters äs they play competitive games. First, they use forms and constructions that belang to the specialized register of sportscasting. Second, a boy acting äs sportscaster takes on a distinctive footing and refers to both himself and his playmate in the thirdperson. In balancing the roles of player and sportscaster, the boys shift footing from moment to moment. They usually step out of the role of sportscaster in order to deal with procedural matters that arise during a game; however, they can also incorporate such events into the sportscasting line.

The X-Bar grammar for stories: Story grammar revisited

Abstract: The Story Grammar (S-G = Story Grammar) approach suggests a model based on the generative framework for representing both story structure and comprehension (cf.,for example, Rumelhart, 1975, 1977; Mandler and Johnson, 1977; Stein and Glenn, 1979; Thorndyke, 1975). The present paper is an attempt to develop the S-G model in two ways: 1. It suggests a radical revision of the S-G model replacing it with an X-Bar story grammar (XBSG), which incorporates the notion ofthe X-Bar grammar äs developed in sentence level grammar (cf. Jackendoff, 1977). I will argue that several fundamental shortcomings of the S-G (among which are the lack of 'descriptive' and Observational' adequacy) can be overcome by a grammar based on the X-Bar model. 2. Ofgreater importance, it attempts tofill in the most significant deficiency of the traditional story grammar s, namely, the lack ofparsing procedures. Based on causal relations between propositions, the parsing procedures which are proposed take äs their input propositions ofthe story andproduce the tree diagram ofthat story äs their Output. Some empirical evidence, based on summary experiments, is introduced in support of the XBSG and the parsing procedures.

On conversational involvement by mail: The use of questions in direct sales letters

Abstract: Numerous studies have focused on written texts and/or spoken language, only to argue eitherfor specific differences between these modalities (for example, Chafe, 1982; Chafe and Danielewicz, 1987) or relative similarities (for example, Tannen, 1982 andelsewhere). Farfewer linguists have analyzed the language used in sales and advertising contexts (for example, Leech, 1966; Geis, 1982), but t o similar end: how language works t o induce voluntary changes in buying behavior remains dimensionally complex and incompletely understood. This paper reports partial results of a study to describe some characteristic features of a previously uninvestigated form of persuasive discourse: direct mail. A corpus consisting of 100 representative (actual) promotional sales letters is examined to determine the extent to which this kind of planned, impersonal marketing communication evidences 'speech like' use of features previously correlated with 'unedited oral genres' (i.e., ordinäry conversation and personal letters; Biber, 1986b). Here, l focus on the use of questions: speech acts which existing research has linked to interactivity and involvement (for example, Tannen, 1984; Lakoff, 1975), 'controV (for example, Goody, 1978; Bogoch and Danet, 1984) and persuasion (for example, Petty et ., 1981, Anzüioti, 1982). Ishow how questions contribuie to the 'conversationality' of direct sales texts. Based on the data, l argue for classing this type of marketing communication äs non-prototypical in its blend of features (following Lakoff, 1982). Finally, I discuss the potential for a düster ofdistinguishing features, and provide preliminary conclusions regarding Strategie employment of'conversational' features in direct mail copy.

Functions of the discourse marker Dakara in Japanese conversation

Abstract: Although the Japanese conjunction däkara is widely recognized äs a logical connector of 'cause-and-result', thispaper argues that däkara has additional discourse functions in casual conversational interaction. Based on examination of 71 tokens of däkara extracted from 20 casual conversations videotaped in Japan, I argue that functions of däkara include marking a point in discourse where supplementary explanation relevant t o the prior text Starts ( sequence-signalling explanatory function). This usage of däkara is significantly frequent (observedin 36.62 per cent of total occurrences of däkara,) in the data examined. Additionally iwo other secondary functions of däkara (äs a marker for [reluctant] repetition, and äs a signal for claiming and ending the turn) are discussed. The multiplicity of functions of däkara is explained by viewing that däkara functions on two different levels of discourse; on the textual semantic level and on the level of interactional move.

Japanese culture versus schizophrenic interpretation

Abstract: The purpose ofthis paper is t o explain Japanese culture by way ofsemiotic typology of Interpretation. I will attempt an explanation of the way in which the Schizophrenie Interpretation of texts is characterized by its weak dependence on context and strong preference for a context-free Interpretation, and that this characteristic is, typologically speaking, necessarily favored to a remarkable extent by technologically developed modern cultures in Gesellschaft society. On the other hand, I will describe how traditional Japanese culture has favored the highly context-dependent Interpretation due to its Gemeinschaft orientation of the social life. Technologically developed Japanese culture, however, necessarily forces Japanese society to Interpret itselfin a context-free oriented way at least formally on the surface, and yet we cannot deny thefact that the Interpretation is still substantially worked out to a great extent according to the traditional ways so deep-rooted in this culture. It may be this heterogeneous context-dependent Interpretation ofthe texts in Japan comparedwith the majority of context-free oriented interpretations in technologically developed modern culture and the coexistence of two apparently contradictory types of Interpretation that tend to give foreign people the impression that Japanese culture is rather difficult to understand. 1. Schizophrenie Interpretation It has been noted through a number of case studies that Schizophrenie tempered people are generally poor at adjusting themselves to accidental, unpredictable, or ambiguous contexts, and that they are usually more sensitive to the subtle delicate signs of coming events well beforehand, This is regarded äs their unconscious effort to turn an uncertain context 01655-4888/89/0009-0351 $2.00 Text 9 (3) (1989) pp. 351-365

Some nominal anaphora in English and Japanese

Abstract: This paper first examines the syntactic and semantic properties of oneanaphora in English and no-anaphora in Japanese. It then attempts to clarify the similarities and differences of English and Japanese nominal anaphora, focusing on the universal restriction on the interpretation of anaphora and the different behavior of the locative phrases with respect to one-anaphora and no-anaphora. This discussion reveals that the wellformedness of Japanese nominal anaphora depends not only on syntactic and semantic structure but also on conceptual structure.

Pragmatic functions of sentence and text coordination in natural language: A case study of Japanese coordinate expressions

Abstract: In coordination, linguistic units to be conjoined are generally supposed to be constituents of t he same type belonging basically to t he same formal category. This might be usually the case when it comes to the coordination ofwords andphrases within a clause or a simplex sentence. In the case ofthe coordination of clauses on intrasentential and intersentential levels, however, the Situation is not that simple. In this paper, I will try to clarify a variety ofways in which various clauses are linked through overt and covert coordination. Especially, an attempt will be made to clarify, among others, the following issues by examining various coordinate expressions in Japanese: (i) Possible clause types that can be coordinated on intrasentential and intersentential levels. (U) Thepragmatic relationship between conjoined clauses and their illocutionary forces. (iii) The difference in speech act function between overt and covert coordinate expressions. (iv) Pragmatic grounds on which the Connectivity and coherence of coordination is made possible. Various pragmatic functions of sentence and text coordination will be made clear by examining Japanese coordinate expressions in connection with these issues.

Idealizations in discourse theory construction: Observational analogues from truthconditional semantics to statutory interpretation

Abstract: This study informally defines 'idealizatiori and specifies three differences in the way in which idealizations are used in empiricist and rationalist linguistic methodologies. These affect the content that specific sets of idealizations should express, the extent to which a sei of idealizations is rendered maximally explicit, and the perceptual acuity ofthe observer the sets are to serve. The study then exemplifies an empiricist methodology in constructing a sei of idealizations for statutory Interpretation. A sei of idealizations underlying a truthconditional linguistic semantic theory is used äs observational analogues. Although many differences distinguish the two Systems, both facilitate a common broad function, the maximally intentionless construal of liieral meaning. Since idealizations work in conjunction with theoretical claims in the case of semantic theory, and wiih various guides to Interpretation in the case of legal studies, one infers correctly that the collective function of the idealizations in both Systems will be to define certain aspects ofa speech event in which the possibility ofmisunderStanding is reducedasfar aspossible. Thus one may discover systematic correlations between the macrofunction of theoretical Statements and the content of auxiliary Statements in the theoretical frarneworks of various disciplines. In Thilosophy and discourse analysis', Kasher (1985) explains that a characteristic philosophical approach is one which critically examines central aspects of a theory in another discipline, focussing particularly on those aspects taken äs assumed, primitive or axiomatic by its members. In reviewing this contribution to the Handbook of Discourse Analysis, Frawley remarks that many discourse analysts would profit if they were to become less content with accepting their own unexamined theoretical notions and were to adopt Kasher's philosophical, critical attitude. NeoEmpiricists, Frawley believes, especially need to clarify their thinking with respect to the role that idealization plays in discourse theory construction; 01655-4888/89/0009-0191 $2.00 Text 9 (2) (1989), pp. 191-229 © Mouton de Gruyter

Text construction in German and Japanese: A comparison

Abstract: A text in Japanese isfirst grasped äs a whole, and then it is subdivided into natural breaks such äs breath-groups. Consequently, there has been no such thing äs the concept of 'sentence* äs a speech unit in the history of Japanese linguistics. This characteristic of the Japanese language certainly results from Japanese mentality and the Japanese patterns of behavior. In this paper the nature of 'so-deixis' in Japanese and of Karl Bühler's parallel 'der-deixis' in German is discussed. Both deixes share the important function of establishing textual coherence, but they differ in their modes of demonstration. While the Japanese approach is 'centripetal', the German approach is 'centrifugal*. Dominant in each case is the cultural principle of Obscuration' in Japanese and of 'clarification' in German in marking the notional boundaries. A problem instead of an introduction 'Language has the tendency to retain its creative power by working in resistance against given conditions.' This Statement by Koojiro Yoshikawa (1952), a renowned Japanese sinologist, seems to find corroboration whereever the dynamic side of language is involved. The same idea is focussed upon by Jan Mukafovsky's characterization of the aesthetic function of language äs a '... dialectic nagation of each practical function' (1967: 49; our translation). It is not easy, however, to divide practical from aesthetic function, or, in Yoshikawa's words, to find out the give conditions of the language in question and at the same time to discern those features that are adding up to the 'resistance' mentioned above. For language not only appears multilayered, but also presents its faceted structure — äs Karl Bühler (1965) points out with profound understand01655-4888/89/0009-0339 $2.00 Text 9 (3) (1989) pp. 339-350 © Mouton de Gruyter

Echo questions in German and Japanese

Abstract: Although t he echo question is restricted in its use in the sense t hat it has to repeat pari of the preceding utterance, it can at the same time contain certain constituents which are not likely to be found in the other types of questions. This latter point, however, has sofar attracted little attention of the researches, who are generally interested in defining a common logicosemantic structure for all the types of question. The present paper proposes to examine such problems äs the relationship between the echo question and direct and indirect speech and the characteristic properties of the echo question äs contrasted with the yes-no question, the -question and the whquestion. In the concluding section, afunctional classification is offered on the basis of the relationship between the echo question and its answer.

On so-called Zero-pronominal: Deletion and host parasite structure

Abstract: In Japanese, grammatically essential sentence elements are quite offen deleted. Deletion is a universal phenomenon which occurs not only in Japanese but in English and other languages. It is supposed that communicatively redundant sentence elements are deleted äs much äs possible. The possibility ofdeletion is, however, differentfrom language to language. The attempts by Kameyama to explain the phenomenon in terms ofthe notion of 'Zero-pronominal' and 'Centering Model' leaves much to be discussed. The aim of this paper is to propose a model, 'host-parasite structure', which expands LFG into the discourse domain, availing itself of 'Thematic Progression Theory'. l. In Japanese sentences, one sees quite often deletions of grammatically essential sentence elements. Kameyama (1985) made an attempt to explicate the deletion of this type in terms of the concept of 'ZeropronominaF (ZPRO) or *Zero-anaphor' (ZA). The point is that the missing noun-phrase were not actually deleted, but expressed by a ZPRO. It is, however, a matter of serious question whether 'the unexpressed' could be 'expressed'. The concept of ZRPO is similar to that of Nullartikel (zeTQ-article). In German, neither the definite nor the indefinite article is used before the indefinite and plural noun. Because no article can be used, Nullartikel is used there. But one is easily misled, by the expression 'use Nullartikel' äs in saying 'use the definite or indefinite article', to think that an article by the name of Nullartikel really exists. To 'use NullartikeF means nothing but to use no article. There exists no other article than the definite and the indefinite articles. ZPRO, too, is the expression of there being no pronoun where it is supposed to be used. If any essential element of a sentence is missing, whether it is the subject or the object, either the element is deleted or the 01655-4888/89/0009-0321 $2.00 Text 9 (3) (1989) pp. 321-338 © Mouton de Gruyter 322 Yutaka Shimokawa sentence is ungrammatical. If the sentence containing ZPRO is not ungrammatical, one will rightly have an intuitive feeling that something is deleted. Deletion is generally treated in the theory of grammar rather äs a marginal matter, and is explained, in English and German, only äs a phenomenon occurring in coordination. The deletion in Japanese is, however, too much of a normal phenomenon to be considered äs a marginal problem. It should be explained, not just äs a stylistic matter, but äs a grammatical problem. In the domain of discourse, deletion is a universal phenomenon occurring not only in Japanese but other languages. The treatment of deletion not äs a phenomenon of an individual language but äs a language universal would make a considerable contribution to the development of the theory of grammar. In this paper, we shall look from the viewpoint of text linguistics into the structure which cannot be explained by the obscure concept of ZPRO, and attempt to find better explanations of it äs a general phenomenon of language. 2.1. ZPRO is explained, before the appearance of Government and Binding Theory (GB Theory), by 'Deletion Transformation' in Generative Transformational Grammar. After the GB Theory, particularly in Lexical Functional Grammar (LFG), ZPRO came to be viewed äs a pronoun with no phonetic content. Kuroda (1965) analyzed the deletion of NP unique to the Japanese sentence äs 'Zero-pronominalization' (ZP), in relation to the Reflexivization (REF), and summarized the conditions of ZP and REF äs follows: (1) When the subject of the constituent sentence is coreferential with the matrix sentence subject, it may either be reflexivized or Zeropronominalized. (2) When the object of the constituent sentence is coreferential with the matrix sentence subject, i t is a. reflexivized if it is dominated by the node VP of the matrix sentence, and b. zero-pronominalized otherwise. That ZP/REF of the coreferential NP is, however, limited to the case where 'Empathy Focus' is placed on the referent of the NP in question was pointed out by Ohso (1976). The conditions of the placement of empathy focus, according to Kuno and Kaburagi (1975), are äs follows: (1) Empathy focus of the Speaker is placed only on one referent in every sentence On so-called Zero-pronominal 323 (2) Empathy focus of the Speaker is likely to be placed on a. the Speaker himself, b. the hearer, or c. the third person referred to by the pronoun, in this order. (3) Empathy focus of the Speaker is more likely to be placed on the referent thematized than on the referent just introduced into the discourse. (4) Empathy focus of the Speaker is more likely to be placed on the referent of the subject than of the object. In the passive construction, empathy focus cannot be placed on the agent. Ohso pointed out that REF and ZP are applied complementarily, the former removing the redundant Information, and the latter preventing ambiguity. On the other hand, while in English the personal pronouns and reflexive pronoun are complementary in their application, in Japanese the NP is sometimes Äare-pronominalized. .Kare-pronominal is the third person pronoun with kare (he), kanozyo (she), and karera (they), and the usage is limited stylistically and sociologically. The conditions of fazre-pronominalization are summarised by Ohso äs follows: When the Speaker is not empathizing with the referent of the subject and when ZP would lead to ambiguity äs to the subject of the relative clause, the relative clause subject can be fcare-pronominalized. Ohso pointed out further the difference that while Equi-NP deletion is applied with a group of verbs which have certain syntactical and semantic characteristics, ZP is applied arbitrarily since it does not depend on the verbs. 2.2. The problem of ZP has come to be treated from a viewpoint basically different from deletion transformation, due to the appearance of LFG which requires no syntactic transformations and treats the grammatical functions äs primitive elements, and to the development of discourse analysis which does not limit its scope to a single sentence. The proposals by Kameyama (1985) to solve the problems of ZPRO will be shown in the following passages. Languages can be divided into four types according to whether major grammatical functions (GFs) with no phonetic content, i.e. ZA, are permitted ( + ZA) or not (-ZA), and whether subject-verb agreement inflection is applied (-l· SV A) or not (SVA); One of the characteristics which the languages of Type 3 have in common is 'rieh semantic specification of arguments', including the 'selectional restrictions', and another is the property of 'topic-prominency', the topic playing a role no less important than the subject. These two character324 Yutaka Shimokawa Table 1. Types of Languages 1. [-ZA, + SVA]: English, German, etc. 2. [4-ZA, +SVA]: Latin, Italian, etc. 3. [ + ZA, -SVA]: Japanese, Chinese, etc. 4. [-ZA, -SVA]: Scandinavians, etc. istics are indeed concerned with the missing of anaphor, but do not explain it efficiently enough. The missing subjects and objects of Type 3 languages are not syntactically recoverable, but pragmatically, i.e. on the basis of verbal or non-verbal contextual elements. Thus discourse-level grammar is essential for the languages of this type. LFG makes it possible to identify the missing element on the basis of Information äs to obligatory arguments stated in the lexicon (Kameyama 1985: 31-53). IN LFG, the sentence structure is represented in two levels: the 'constituent structure' (c-structure), representing the surface configuration of word and phrases in the sentence and the 'functional structure' (fstructure), representing the underlying structure of the sentence. And the GFs subcategorizing verbs are divided into 'semantically unrestricted GFs', such äs subject (SUBJ), direct object (OBJ), and indirect object (OBJ2), and 'semantically restricted GFs', such äs oblique object (OBL), complement (COMP), and complement without the subject (XCOMP). ZA does not appear on the c-structure, but while semantic ZA, i.e. ZA of semantically restricted GFs, does not appear even on the f-structure, syntactic ZA, i.e. ZA of semantically unrestricted GFs, appears on the fstructure. Consequently, only the syntactic ZA is treated äs 'Zeropronominal' (Kameyama, 1985: 61-88). An utterance contains some referring expressions which help to identify each non-linguistic entity called a 'center'. The entity 'centrally talked about' is called 'backward-looking center' (Cb) or just a 'Center', and the entity which is possible to pay attention to 'forward-looking center' (Cf)· An utterance is associated with one or zero Cb and any number of Cfs. Cfs are ordered according to 'how much a Cf is expected to be the next Cb' äs shown in Table 2: The initial utterance of a discourse unit has no Cb, and is followed by any number (zero or more) of utterances with the same (or a closely related) Cb retained until an utterance with no Cb appears. This process is called Table 2. The expected center order in Japanese 1. TOPIC (SUBJ > OBJ > OBJ2 > POSSESSOR) 2, SUBJ 3. OBJ >OBJ2 4. others (Kameyama, 1985: 115) On so-called Zero-pronominal 325 'Centering', and the minimal discourse unit is called a 'Centering Unit'. Regarding a Cb äs an entity identified by ZPRO makes it needless to describe ZPRO äs a 'deleted overt NP', and the 'referential property and discourse function' peculiar to ZPRO become explainable. (Kameyama, 1985: 74). 3.1. While Ohso (1976) pointed out that the usage of kare-pronominals is stylistically and sociologically limited, Kameyama (1985) overlooked this fact. The result is that she invented a concept of so-called *fprecedence', äs if in LFG the order of elements which is the relation on the c-structure, could influence anaphoric control. She, by doing so, distorted the principle of LFG. (1) [Kare??ilj no : (eifj)] kuruma de Taro{ ga kaigan ni he of car by Taro NOM coast to itta (koto) went that (III70a) '(that