1. Introduction

Concern about the written production of university students has become a recurring theme in publications and presentations in different forums for studies of language and education most recently. This is due to observations in several studies of the low quality of texts in relation to the language skill required at this level of education. The problems range from the acquaintance of simple rules of syntactic constructions and lexical-grammatical uses to difficulties of understanding and organizing ideas and information, along with a lack of knowledge of textual organization principles and other problems.

One of the explanations for this situation is the precarious formation these students had in previous levels of education, but in fact, it is worse within the universities. Contrary to expectations, students’ access to different textual genres produced in the academic community is often fragmented and limited to material they have copied.

University programs have not properly invested in teaching textual genres, ability considered necessary for preparing students in the academic community to make decisions in contemporary society.

In the academic community, "Genre analysis as an insightful and thick description of academic and professional texts has become a powerful and useful tool to arrive at significant form-function correlations which can be utilized for a number of applied linguistic purposes (…)” [1].

Though genre studies have been raised in its initial phase to make a classification of what seemed to be a confusion of texts, they were primarily consolidated because the traditional approach could not deal with situations where individuals actually use the language as a tool for interaction, reproduction or social change. More than the formal aspects of language - proper Structuralism - and going beyond the study of linguistic, pragmatic and interactional factors represented by Textual linguistics, the main concern of genres analysis is with the texts in use: written by whom, for what, to whom, how, under what circumstances, with what degree of transparency or hegemonic and ideological camouflage [2].

This problem is the motivation for the present study. In undergraduate courses at a public federal university, we have observed texts production that confirm the issues already raised by other studies of a similar nature. Among the existing genres in the university, we found a significant presence of abstracts in different areas, but a preliminary search indicated a higher concentration in the area of Social and Human Sciences. This fact justifies, in part, the choice of the theme of this research, but two other factors may be considered as well: 1) these textual genres, in general, are synthetic, selective in presenting the ideas of a text and condensing the content of a reading in order to give an overview of the original text; 2) through these texts it becomes possible to evaluate, to some extent, the understanding of the original text.

In the study of abstracts, we identify and analyze some regular characteristics of the genre which organizes, establishes and determines the textual genres. The line of analysis that guided this research is based mainly on studies developed by Miller [3, 4] in the "New Rhetoric", i.e., the recognition of rhetorical dimensions in knowledge construction; the research developed by Swales [5] and Bhatia [6, 7] on rhetorical movements of textual genres; the principles of understanding the texts presented by van Dijk and Kintsch [8]; the important contribution of Bakhtin [9, 10] to studies of genres and language; and the latest discourse sociointeracionist approaches developed by Bronckart [11] and others. The emphasis in this study focuses on the rhetorical-discursive actions of textual organization, which determine macro and microdiscursive options to be made by the producers of texts.

In this perspective, the analysis of abstracts requires considering not only enunciative context of production and reception, but also rhetoric organization, its textual aspects and rhetorical discoursive devices, and emphasizing the main characteristics and linguistic marks of texts within the social practices in which they take place.

To develop this study we used data from a survey conducted the first semester of 2009 in classes of undergraduate students who answered a questionnaire, the aim of which was to gather information about options and frequency of reading and writing.

In the next phase we collected 167 abstracts of undergraduate students of Humanities and Social courses from two universities. Selected based on general criteria, such as presentation of references, length, textual organization and possibility of consulting the original text (see Table 1) we examined summaries of 45 students in graduate programs in Arts, Administration, History and Sociology courses, as it can be seen in Table 2.

Table 1. General criteria for abstracts’ selection

Table 2. Original texts and abstracts

The rhetorical-discursive elements selected as categories for analysis are briefly presented below:

a) the socio-interactionist postulates presented by Bakhtin and developed by the School of Geneva’s researchers, especially Dolz [12], Schneuwly [13, 14] and Bronckart [11]; and

b) the studies developed by van Dijk and Kintsch [8] on the organizational principles of complex information processing, i.e. general macrorules of text understanding, which, in the case of abstracts, apply for the reduction of relevant semantic information.

To illustrate these categories we present parts of the abstracts in which there have been observed (or not) these occurrences. In order to make the recognition of these aspects more accessible, this analysis has been divided into two topics: 1) metadiscursive markers; 2) macrorules of texts understanding.

2. Metadiscursive Markers

Based on the model developed by Motta-Roth et al [15] for abstracts analysis, we considered the following markers.

2.1. Text Structure Markers (TSM)

Indicates how different blocks of information are organized and places the reader in the summarized text. They can be markers of continuity (first, the, in third place) and time connectors (at the same time, then again). The following examples illustrate this item:

(1) From this point on, it was found that many tourists, including foreigners, have contributed to environmental degradation ... (SOC/5)

(2) In summary, it was examined the gradual progression of environmental impacts ... (SOC/6)

(3) Therefore, considering what was found, the data showed a gradual worsening of the environmental impacts ... (HT/1)

(4) Thus, there are three classic positions with respect to the subject ... (LT/3)

(5) Koch, below, shows us that the definitions depend on how the subject is regarded. (LT/14)

(6) And as said before, the strategy of the brand is one of the triumphs of the company ... (ADM/1)

(7) First there is the tendency of replacing old equipment with new ones (...). Secondly, the high cost of licenses for 3G technology puts pressure on telephone operators so that they start launching new services. (ADM/3)

Text Structure Markers (TSM) in the segments of abstracts analyzed indicate a textual progression, i.e. sets the linear structure of the text, organizing it in order to facilitate interpretive processing, or, as explained by Maingueneau, designating these as "markers of linear integration". Koch [16] notes that these markers play the function of opening, intermediating and closing in the spatial organization of the text. However, their absence does not necessarily indicate that a text does not display this progression. Other elements in the text ensure its linearity. In this case, some factors such as the textual cohesion or other linguistic procedures can ensure the continuation of the theme, as seen in the following fragment:

(8) Defining the subject depends on how language is considered, and from this varied concept, one can choose the most appropriate concept of that. Consequently, the text is in accordance with these two prior terms to, thereby, acquire the appropriate meaning for each type of relationship.

If language is a representation of thought, it is seen as an instrument and its producer is responsible for the meaning (conscious and historical human being). The interactant - reader / listener - is nothing more than a passive being that can discover the mental intention (meaning) of the author within its logical product (text). (...) (LT/6)

According to Biasi-Rodrigues [17], the information in abstracts is distributed in cells or in short thematic groups which do not always connect to each other by means of explicit cohesive ties. Often the thematic sequence is given by the relationship of lexical association, which is anchored in knowledge of the world of academic discourse community.

However, as shown in example (8), you cannot identify textual genre as an abstract with the absence of these markers. In this case there was not even the presence of introduction elements that could enable the genre schema as a whole. Therefore, we say that, in addition to expressions like "the author of the text," "this or that research", "the data analyzed by the researcher", etc., the structural markers play a significant role in the recognition of the abstract genre.

2.2. Attitude Markers (AM)

Show the position of the author of the abstract, both in respect of the author of the original text as well as the contents of it. Take the examples:

(9) The article brings interesting features ... (ADM/1)

(10) Despite the size and scale of the organization one can see the importance of analysis of situations ... (ADM/4)

(11) ... saga was shown (briefly reported) ... (ADM/5)

(12) I agree with Dascal (LT/2)

(13) Text rich in information very enlightening on the issues in question, all covered by a easy language. (LT/5)

(14) The concept of meaning is handled through some theories that, perhaps for brevity of explanation, have not become very clear for me. (LT/14)

(15) ... I believe that different forms of approach should not be placed in extreme opposition ... (LT/14)

(16) ... Antonio Houaiss addresses the teaching of standard language, claiming the reader to be aware of its importance. (LT/19)

The evaluative attitude facing material to be reduced has been widely disseminated among genres analysts as preferably limited to reviews (or critical reviews). Terms of praise and criticism, or "non-specific terms”, which tell the reader how the messages should be interpreted [18, 19] and comparisons are elements that characterize the review. Moreover, one of the classic recommendations for the production of abstracts is that the language to be used in this genre is "objective, clear, and formal". Such recommendations may seem incompatible with the presence of evaluative elements in abstracts and, in fact, appear very discreetly in our corpus. In segments (9) to (16) we observe that, in general, these events do not contradict the principles of objectivity, clarity and formality required by the genre.

In his formulations on the issue of genres, Bakhtin [9, 10] is emphatic in considering the statement as reflecting the individuality of the person who speaks or writes which sets the style. Thus, "... the selection of a grammatical style the speaker makes is an act" [10]. In the case of the fragments we analyzed, the low incidence of these markers may reveal the gap between author and reader, which contradicts the dialogical principle of discourse developed by Bakhtin [10] for whom: "The substantial index (constituent) of the statement is to be directed to the addressee."

Koch called these attitude markers "attitudinal or affective" and "perform the psychological attitude with which the speaker is represented before the events in the statement" [16]. Segments (9), (10), (12), (13), (14) and (15) illustrate this point. Examples (11), (13), (14) and (16) mark the position of the abstract on how the author of the original text developed his ideas. Thus, in these examples the evaluation is made in the two ways previously mentioned: on the content and on the form of the original text.

Therefore, if in theory it is recommended that the abstracts be objective, in the sense that there must not be value-judgment, the presence of certain verbs – especially those of elocution – show a position or attitude by the author of the abstract which may be speculative, presumptive or prospective. Examples (12) and (15) illustrate this position evaluation.

2.3. Validity Markers (VM)

Indicate the extent to which the author believes the propositional content of the text is true, as we see in the following examples:

(17) Considering data that were presented during the research we come to the conclusion ... (SOC/2)

(18) Considering what was found we came to the conclusion... (SOC/4)

(19) These data demonstrate the need for development of sustainable tourism ... (HT/2)

(20) ... Dascal considers himself follower of pragmatic model ... (LT/2)

(21) The author initially considers two factors ... (LT/17)

(22) Although, for the author, the reality is different. (ADM/3)

(23) How has the actual text, "The senior management time devoted to the development of internal values of the company (...)" (ADM/F/5)

It is important to mention the audience – real or virtual, concrete or abstract, identifiable or not – that guides the textual production. The audience is defined as a starting point for implementation of the text, which is to Hoey [20].

The audience of a text is the intended readership, the imaginary person or persons whom the writer addresses and whose questions s/he tries to answer. Ultimately the audience is always a figment of the writer’s imagination since no writer, however skilled, can ever get inside someone else’s mind so completely as to know exactly what they want and need to learn. In composing his/her text, the writer makes assumptions about the state of knowledge of the audience and these are reflected in his/her grammar in quite subtle ways (p.14).

One of these ways is the presence of markers that indicate the dialogic relationship between author and reader, with the mediation of the text that is – in our case – the text to be summarized.

The use of such elements, especially as we see in examples (20), (21), (22) and (23), shows that the author of the abstract, when evaluating the thematic content of the text, introduces other voices that take the responsibility for what is stated. While they state their positions, we realize that the authors in the examples above limit themselves in these evaluations when they go to the original texts in a signal of agreement or disagreement with its composition.

3. Macrorules of Comprehension and Reduction of Information in Texts

In order to understand the problem of that specific genre’s production, many scholars studied the strategies of the language used in abstracts and their basic features. Many of these works are based on studies of Kintsch and van Dijk [8], which propose a model for an abstracts’ production from the premise that the information to be included in an abstract is defined by macrorules (deletion, generalization and substitution) operating in the text-based proposition to produce a macrostructure.

In several studies performed mainly in the 1970’s [8, 21-23], the authors emphasize their concern in elucidating "the rules and principles of reduction of complex semantic information in general, and the relationship between microstructure and macrostructure” [8]. Thus they propose a theory that provides a partial explanation of classical concepts such as those of schemas and frames.

In these studies the authors' attention turns to the semantic structure of discourse, i.e. the abstract logical structure underlying the discourse, and they neglect morphological and syntactic surface structures. One of the central notions of this theory is that the macrostructure – global meaning structure of a text – that composes the basis of a text must satisfy two principles:

1) Principle of involvement - each macrostructure must be involved (induced) by the structure of which it is derived. A macrostructure derives from the text, but not in a deductive way, i.e., a macrostructure is also an inductive interpretation of a text.

2) Principle of relative pertinence - propositions that are of another macrostructure’s presupposition cannot be deleted.

Brown & Day [24] declare that “the ability to summarize information is an important study skill involving both comprehension of, and attention to, importance at the expense of trivia.” In the construction of the understanding process they include, among other skills, disregarding the explicit propositions of the text, inferring the implicit ones, and connecting these propositions in order to obtain the textual macrostructure, i.e. the global coherence of the text got when some strategies or rules for this purpose are trigged. Such rules, which Van Dijk and Kintsch [8] called macrorules, ensure the textual organization and are important in reducing information in order to form a new textual genre, the abstract, although they have not been stated in the context of a theory of genres.

As we propose to learn how students reduce, reproduce or distort the original text in the production of their summaries, we consider in our analysis the macrorules of selection/deletion and integration/construction. Note that the first pair comprises a rule of selection of information to be used in the abstract, as deleting, in fact, is the result of the act of choosing or selecting, while the other pair can be considered a rule of substitution. The similarities and differences between these rules will be properly explained when we present the examples.

a) selection/deletion - It is the ability the abstract producer must show of discarding material that is considered irrelevant and to avoid unnecessary repetition. One must be careful not to discard items that are not a condition of another statement or interpretation of the text as a whole.

Most of the examined abstracts presented irrelevant information that did not contribute to the understanding of the text, making these texts longer than recommended. One that drew our attention was the use of numerous references of research in the abstract. Paradoxically, we had to discard a number of abstracts that had not provided references of the original text. Many simply copied bibliographic data provided in the references of articles read as shown in the following examples:

24) This research is based on authors such as: PINTO, T. K. Structure of the Micro Fauna of the Bank of Coroa do Avião, Recife, Taciana Pinto Kramer, 1998, pg. 51. - LIRA, L. Channel dredging in Gavoa. Environmental Assessment of Ecosystem Itamaracá, Pernambuco. Recife, 1992, pg. 35-41: Environment of the Coroa do Avião. (SOC/3).

(25) This research was based on texts by authors such as T. K. Pinto, L. Lira, L. S. Andrade, G. F. Dias, R. Regner, M. L. S. Tales, C. R. Campelo and E. M. Oliveira. (HT/1)

Pinto [25] observes that "identifying the main idea becomes even more complicated by the fact that, in the same text, this can vary, not only by the readers but also by the situation, depending on the purpose of reading". On this issue the author also points out that the intentions of the reader are as important to understanding as the author’s. Understanding these intentions and distinguishing between textual important information (established by the author as he organizes his ideas) and contextual important information (captured by the reader from his prior knowledge) has been the major challenge faced by producers of abstracts. Indeed, it is possible to say that the good reader is evaluated by the ability to identify and use relevant information, which requires knowledge of the principles of discourse organization submitted, according to Pinto [25], to the following conditions:

1. Recognition of the elements of the textual structure: the context (cohesive and cognitive elements) which provides the content of the text, the internal structure, and the identification of the purposes of the author [26] besides facilitating the understanding of the superstructure.

2. Explicit textual evidence or relevant signs [23] - graphic, syntactic, semantic, schematic and super structural evidence.

3. Prior knowledge - acquired through exposure to tasks in successive readings.

b) integration/construction – in this macrorule a set of clauses may be replaced by a new clause in which all the others are included. In this case, the process to be considered is the integration of information in a single, more central one. In some studies this rule is called substitution.

To illustrate this macrorule, we selected a paragraph from one of the original texts and presented some excerpts that refer to this paragraph to evaluate to what extent the authors of the abstracts were able to identify the central idea and how this is presented in their texts.

"The high board has devoted a long time to the development of internal values of the company, which soon became the basis for brand strategy. Nokia has realized that the consumer is afraid of the world dominated by technology and emphasized the human being” (Alonso, 2001:34).

In the following examples (26) and (27), the literal transcription of part of the paragraph is eloquent enough, as were other examples of this nature.

(26) The high board has devoted a long time to the development of internal values of the company, which soon became the basis for brand strategy (...) (ADM/F/1)

(27) " The high board has devoted to the development of internal values of the company, which soon became the basis for brand strategy. Nokia has realized that the consumer is afraid of the world dominated by technology and emphasized the human being" (ADM/F/5).

Another complex point is the ability to find relationships between ideas, as needed to understand the macrostructure of the text. Considering this point, many scholars note in their analysis that the university students do not prepare a plan in advance to integrate the organization of the text and its content, tending to focus on the first paragraphs of the text or on information that they considered interesting. Several studies have indicated that these students produce long abstracts, similar to a list of propositions without connection, and demonstrate low knowledge of how form and content are related to the purposes of understanding and producing texts.

4. Conclusions

Among the abstract criteria chosen to study at the university, the least problematic seems to be the use of metadiscoursive markers. In general the analysis indicated that the texts produced by the students in graduate programs are far from meeting the requirements of the academic community – some very simply because they have not internalized the idea that communicative purpose, audience and context affect the production of texts.

Observing these students’ testimonies and especially their production, the conclusion that very little has been done to resolve this issue is inevitable. Teacher and students communication seem distant from each other, and the latter lose space because they know little about the rules of production of texts that circulate in this discourse community. To fight these difficulties in the process of writing texts, teachers cannot disregard the importance of genres teaching in institutional contexts, such as the university, so that students will more likely build their identity in the text as much as they articulate their language, and their speech will be more legitimate with the better management of the sociolinguistics conventions of the community.

Insofar as the production of the students is understood not as a product in itself but as a result of socio-historically situated language activity, they will start their work by questioning: what is the goal of my writing? Who is my reader? Where will my text be circulated? Under what conditions will it be read? How will it be evaluated? Thus, in this context they will become part of this complex and diverse academic community.