Coreferential elements, co-reference relations
(Analysis of Hungarian language texts)
0. Officina Textologica, a new periodical by the Publishing House of Lajos Kossuth University of Debrecen, has been originated in order to contribute to the elaboration of a text linguistic - textological method for analysing text sentences. (As regards the main functions of this periodical, see its first issue: János S. Petõfi, A polyglot research program in textology / text linguistics. Officina Textologica 1, 1997.)
As far as linguistics is concerned, we should draw a clear distinction between the so-called architectonics-oriented linguistics and other branches of linguistics (e.g. psycholinguistics, ethnolinguistics, sociolinguistics) on the one hand; and, within the scope of architectonics-oriented linguistics, we should separate system linguistics and text linguistics on the other hand. Moreover, system linguistics should be divided into two components: the relational component and the transformational component, respectively. The system of linguistic disciplines outlined above can be illustrated in a tree-like diagram as shown in Fig 1:
linguistic disciplines
architectonics-oriented linguistics other branches of linguistics
system linguistics text linguistics
relational component transformational component
(Fig. 1)
1. Architectonics-oriented linguistics deals with the lexicogrammatical structure and formation, as well as prosodic, metric and rhythmic characteristics of verbal units.
- The relational component of system linguistics, studying the construction of basic lexicogrammatical structures as elements of the language system, elaborates methods for representing the existing lexicogrammatical relations between components of these lexicogrammatical structures.
- The transformational component of system linguistics studies all possible linear manifestations of basic lexicogrammatical structures within a specific language, and attempts to outline the conditions under which a certain manifestation could be applied.
- The function of text linguistics is, on the one hand, to assign a 'basic lexicogrammatical structure' - representation (hereafter referred to as BLS-representation) to the text sentences of those texts that are being examined, and to provide information about all possible linear arrangements of basic lexicogrammatical structures having been assigned to text sentences. On the other hand, another function of text linguistics is to present arguments for explaining the occurrences of the specific manifestations of basic lexicogrammatical structures that have been actually applied in different places of texts.
Since text sentences seem to be incomplete in most cases, that is, they cannot be considered, as opposed to system sentences, to be independent of each other, it is necessary to transform text sentences into something like system sentences by completing them before the assignment of BLS-representation. When completing text sentences and presenting arguments about their manifestations, we need, of course, competence in linguistics; however, in most cases, it is essential to have special knowledge about textology, and general knowledge about 'the world-fragments' that can be associated with all the subjects concerned.
When analysing a selected text the following steps should be executed in order that the lexicogrammatical structure and formation of each text sentence can be interpreted:
(1) the text sentence should be completed with all the necessary information that can be acquired from previous text sentences;
(2) a BLS-representation should be assigned to the completed version of the text sentence; then
(3) the BLS-representation of Step 2 should be provided with information about its possible linear manifestations; and finally,
(4) arguments are to be sought that can explain
(a) why a specific linear manifestation is applied in a given place of text; and
(b) why and how this linear manifestation can be deduced by transformation from the corresponding of Step 2 (by means of deleting some parts of the BLS-representation or replacing them with pronouns) taking into account whether the text sentence in question is complete or incomplete.
Both in Step 1 and Step 4, it is important to consider co-reference relations between text sentences. That is the reason why we need general knowledge for executing the process, beyond the necessary competence in linguistics.
2. The four steps described above can be considered as text linguistic operations. By them, the conceptional framework of our research program in textology / text linguistics can be further specified as a framework covering all aspects of architectonics-oriented (text) linguistics. The thematical volumes of Officina Textologica, which will contain studies on architectonics-oriented (text) linguistics, will be in accordance with this framework. (As for the conceptual framework, see the first issue of Officina Textologica.)
The main function of architectonics-oriented linguistics is the analysis of text sentences. The planned thematic volumes of Officina Textologica, in which the main issues of the analysis of text sentences are going to be discussed in order that these volumes should also be used as textbooks, can be classified into the following groups (the order of items corresponds to the logic of the research strategy):
(a) completing the text sentences of various texts with verbal elements and reference or co-reference indices;
(b) analysing the possible linear arrangements of text sentences of various texts;
(c) assigning BLS-representations as well as information on their possible linear manifestations, to the completed text sentences of texts;
(d) generating text sentences from the BLS-representations on the basis of the information having been assigned to the representations.
3. The aims of Vol. 2. of Officina Textologica can be declared as follows:
(1) to show how reference is made, in different text types, to the elements (e.g. persons, objects, events, etc.) of a specific part of the world ('world entity'), which is presumably represented in a selected text, according to the sender (author, speaker, etc.) and/or the receiver (reader, listener, etc.) of the text; in other words, what sort of words and/or suffixes can be used as referential units to the 'world entities'?
(2) to show what kind of linguistic, textological, or general knowledge (about the World, so to say) is needed for revealing the co-reference relations between coreferential words or phrases referring to the same 'world entity';
(3) to search for the most effective way of representing and illustrating the coreferential elements and co-reference relations.
This current volume of Officina Textologica belongs to Group (a) within the classification listed in Section 2. above.
The authors of this volume are scholars most of them teaching Hungarian language and/or literature at universities or colleges. They have been invited to recommend two or three texts for co-reference analysis. The texts, one text for each scholar so as to be actually analysed, had been selected by the editor-in-chief, taking good care of they belonging to different type of texts. Besides, for the sake of consistency and that the methods could be compared directly with each other, or, in other words, could be projected onto each other, an 'orientating analysis' (in Study 1., see below) had been made available to the scholars in addition to the above mentioned reference to Vol. 1 of Officina Textologica. This second volume collects together the analyses of those texts selected, and contains a short introduction, and a longer epilogue as well.
Study 1, the author of which is JÁNOS S. PETÕFI, editor-in-chief of Officina Textologica, contains a co-reference analysis of a biblical text (Mt. 9:9-13, The Calling of Matthew), a parable extracted from New Testament. This is the above mentioned 'orientating analysis' which presents the suggested steps of an effective co-reference analysis. The author takes good care of demonstrating co-reference relations within the text (or text sentences) explicitly.
Study 2 deals with an extract from the Revelation (Rev 21:9-23, from The New Jerusalem). The authors, JUDIT PORKOLÁB and KÁROLY I. BODA, at first analyse the text in a broader framework, following an approach concerning the theory of literature. The possible interpretation of symbols, as well as the various associative links relating the synthesis of the books in Old Testament and the books in New Testament, are then interpreted by analysing the coreferential elements and phrases, and co-reference relations in the text (or text sentences). The result is illustrated in a matrix invented by the authors, representing the co-reference relations between text sentences by the number of occurrences of co-reference indices in each cell of the matrix.
Study 3 consists of a co-reference analysis of an encyclopaedia entry (Hungarian Larousse Encyclopaedia, 'chariots race'). The author, EDIT DOBI, adapting the method described in the 'orientating analysis' to the analysis of the given encyclopaedia entry, attempts to answer some special questions and problems regarding this type of text, e.g. the separation problem of text sentences, the interpretation of the various signs and abbreviations used in the encyclopaedia, etc. Moreover, the author illustrates the concept of 'partial co-reference', and some interesting metalinguistic phenomena which follow the definitive nature of encyclopaedia entries.
Study 4 is a co-reference analysis of a selected recipe (from Lukács Túrós: Cookery Book for Maidens and Women, 'chicken with rice'). The analysis of text sentences is based here on the 'orientating analysis', too, but the author, ÁGNES DOMONKOSI, raises some further questions about, and succeeds in finding solutions to, the representation of some text-specific problems, e.g. the alteration in the referential value or strength of some dominant language elements in the continuous process of cooking, the typographical and graphic representation of recipes, etc. The latter brings up some essential questions concerning co-reference analysis, and might as well need further discussion.
In Study 5 the author, ERZSÉBET CS. JÓNÁS, analyses the coreferential elements and their relations in a business letter written for advertising purposes. The author, taking into consideration some sociological factors, studies the verbal manifestations of the communicative functions of this type of text, which all reflect the main purpose of convincing the reader. Besides, she attempts to determine what knowledge (about the language system, the use of its elements, or a virtual 'world entity') is needed for the proper description of the structure and formation of, as well as co-reference relations, 'coreferentiality' within the text (or text sentences).
Study 6 contains the co-reference analysis of a text written in specialist language (Elemér Hankiss: Literary work as complex model, the first two paragraphs in the first chapter of the book). The study follows exactly the 'orientating analysis' during the coreferential description process, intending to illustrate the genre-specific characteristics of the text. Considering the stronger and weaker relations within the series of co-reference relations, 'co-reference chains', the author, FRANCISKA SKUTTA, doubts, in some cases, the full referential equivalence of certain language elements when the are in co-reference relations with each other. For the illustration of that, she suggests that new type of co-reference indices should be introduced.
In Study 7, the author analyses the co-reference relations of a narrative literary text extracted from a Palóc (Hungarian) folk-tale. MÁRTA TUBA adopts the methods and system of notation described in the 'orientating analysis' and adds new elements to them, which seem to be very effective for the representation of co-reference relations, when illustrating certain characteristics of this type of text. According to the author, the series of events happened one after the other plays a key role in the organisation of this type of text; the author therefore attempts to illuminate the coreferential aspects of the means of expression relating to the elapsing time in tales.
The second part of Vol. 2 of Officina Textologica begins with Study 8. The authors, LÁSZLÓ VASS and RÉKA BENKES, analyse, within the framework of semiotic textology, the coreferential organisation of a multimedia product (Hungarian Bookshelf CD-ROM, its Help feature). The authors show that the CD-ROM consists of verbal components as well as images, and its Help feature can be considered as a specifically organised, non-linear type of text, a so-called hypertext, which needs, of course, a linear manifestation when used as 'directions for use'. According to the authors, the co-reference analysis of this medium requires some interpolations, sometimes implicitly, in the linear manifestation of hypertext. These interpolations can be obtained directly from the mental model of this manifestation (or its relational structure) on the one hand; and experimentally on the other hand, when one actually being confronted with this type of communicative means. The authors attempt to explore these interpolations, and solve the problem of explicit representation of (co)-reference relations using various types of referential or coreferential indices.
The author of Study 9, IMRE BÉKÉSI, analyses so-called `key sentences' extracted from a novel (L. Tolstoy, Anna Karenina, first part, XII, 3-4.), with special respect to their role in the organisation of texts. The author contributes to the theoretic foundation of co-reference analysis by stating that the dominant lexical components of `the key sentences' could be the core of 'co-reference chains' that organise the text.
The author of Study 10, KATALIN B. FEJES, analyses the co-reference relations of a poem (Attila József, The Greeting of Thomas Mann). The author demonstrates the role of the 'co-reference chains' that correspond to different and sometimes alternating ways of communication (such as auditory, visual, and kinetic channels) in the organisation of texts, taking into consideration the imaginary situation the poem depicts (i.e. the greeting of Thomas Mann by the poet himself). This situation, 'the birth of the poem', could have aroused so strong emotions 'that influenced the whole composition of the poem', the author says.
In Study 11, the author, IRMA SZIKSZAI NAGY, analyses the co-reference relations of a newspaper article (Kurír, VII (1997) 5, p. 16, 'Bike-party leader' (approx.)) with a slightly different approach, compared with that in the 'orientating analysis', to 'coreferentiality'. Taking into account all the phrases that are in co-reference relations with each other as well as the co-reference relations themselves, the author investigates the so-called 'clusters' of coreferential elements (counting up the number of occurrences and relative frequencies of the elements), and examines the occurrence of coreferential elements in macro-structure units, their 'expanding effect'. Besides, the author illustrates the 'co-reference chains' of the dominant language elements, which play an important role in the organisation of text, and shows the stylistic value of 'co-reference chains' by considering the keywords that can be substituted with each other.
Study 12 contains a co-reference analysis of an 'article' selected from a so-called dictionary-novel (Ferenc Temesi: Por. eszménykép.). It is based on the 'orientating analysis' but, as the author, GÁBOR TOLCSVAI NAGY declares, the interpolations and explanations seem to be less defined than those in 'orientating analysis'. The reason of it, first of all, can be attributed to the peculiarities of the text analysed, which can be considered as both an independent entity, and part of a broader, more complete text.
Summarising our considerations, this second volume of Officina Textologica can hopefully serve as a systematic and thematical collection of problematic issues and examples, covering all main aspects of the subject discussed above. The volume might as well be a starting-point for doing practices in co-reference analysis for educational purposes; besides, it can be a reference book for further researches of the subject..