A Product-Focused Approach to Text Summarisation

Esther Uso Juan
Juan Carlos Palmer Silveira
palmerj [at] mail.uji.es
Universitat Jaume I-Castello, Spain

The authors of this study investigated whether summary writing instructions and second language (L2) proficiency level account for differences in the L2 summary writing performance for two groups of students. The 15 students in group A were instructed in the rules of summary writing and had an intermediate level of English. The 15 students in group B were not instructed in the rules of summary writing and had an advanced level of English. We carried out research based on a comparative product analysis, which focused on four aspects: the quality of the abridgement, the summarising strategies used by the students, the extra-textual information included in the abridgements and the rhetorical structure followed. Results indicated that the fact of having clear instructions regarding what is expected from a summary helped students in group A to enhance their writing ability and perform quite similarly to the students with an advanced English level.

Introduction

The ability to summarise information is an essential skill in University studies. Most students, through their academic life, have to condense information from lectures, journals, textbooks and other bibliographical sources in their disciplines in order to fulfill certain assignments in their own field of study. As Stotesbury (1990: 3) stated, "summarising entails the reduction of a text to its essential constituents which means that students have to be able to grasp the overall structure of a text and be able to distinguish the major issues from the minor ones."

The task of summarising in L2 has began to receive attention in the ESL courses in the last few years. In fact, it is seen as a highly productive task by many ESL instructors because, as Palmer (1996: 123) pointed out, "it implies both the complete comprehension of the text to be abridged and the necessary writing ability to create a new version of the source text." Sarig (1988: 4) considered summarising tasks as "junctions where reading and writing take place." Therefore if we teach our students how to sum up a text, and how to condense its information, we will similarly enhance their reading and writing ability. It seems reasonable to think that our ESL classes can obtain some benefits by the teaching of summarising strategies; helping our students to understand complete sets of information allows them to develop their writing skills. This seems to create a holistic view of language use, where all other aspects will be affected by the use of production and interpretation strategies throughout the use a whole set of procedures. Summary writing, therefore, is not merely a linguistic activity, but also a communicative and discoursive one, in which students apply the knowledge previously acquired.

In the attempt to link reading comprehension and writing fluency, summarising is also a very motivating teaching task. This type of activity implies the use of diverse cognitive mechanisms, as many linguists and psychologists have commented in the past (i.e., van Dijk, 1977a, 1977b, 1979; Bracewell, 1981; Flower & Hayes, 1981; Hayes and Flower, 1983; Hidi and Anderson, 1986; Amlund, Kardash, & Kulhavy, 1986; Carey & Fowler, 1986; Eysenck & Keane, 1990). Creating a summary is also identified with the development of organizing abilities, based on the rearrangement of the information in a way that should be both clear and appealing (Donin, Bracewell, Frederiksen, and Dillinger, 1992). Finally this type of activity also shows the full comprehension of the source text; only those who can understand the original piece of discourse will be able to create an abridgement by condensing the gist of that first text (Winograd, 1984; Stotesbury, 1990).

Nevertheless, much of the research about summarising has focused mainly on L1, taking into account both process and product (Kintsch & van Dijk, 1978: Brown and Day, 1983; Winograd, 1984); on the contrary, only a reduced number of studies have focused on L2, devoting their attention to the different ways summarising strategies can enhance the students« performance in the ESP classroom context (Stotesbury, 1990; Palmer 1997, forthcoming).

Considering the literature on text summarising, there seems to be an agreement on the fact that this is a difficult task. In order to master this skill, students need to understand discourse perfectly and be able to carry out the necessary operations so as to arrive at the gist of the information (Stotesbury 1990). Aiming at helping our students to perform better when abridging a text in an L2, as well as trying to facilitate this difficult task, we decided to carry out research based on a comparative product analysis. Bearing in mind this objective, we have observed four different aspects on the students summaries:

Quality of the summary
Summarising strategies used
The role of extra-textual information
The rhetorical structure followed by the students

Method

Subjects

Two different groups of Spanish students were engaged in the experiment as part of their class requirement. The first group (henceforth known as group A) was made up of fifteen Spanish first-year university students majoring in Education. Their ages ranged from 18 to 21 years. They had studied English for an average of 7 years, mainly through formal education in Spain. Previous tests confirmed that their English proficiency level could be defined as intermediate.

The second group analysed (henceforth described as group B) was made up of Spanish teachers of English, all of them working in primary schools, who were attending an advanced course in English. Their ages ranged from 25 to 40 years. All had studied English for the usual 7 years at secondary school, plus three additional years at University. All these subjects were currently teaching English at primary schools in the Castell—n area. A number of tests confirmed that all these subjects had an advanced English proficiency level.

In order to corroborate their level in English, these two sets of learners with different L2 proficiency levels were asked to complete an additional placement test, which was administered at the beginning of the course. The test consisted of 50 multiple-choice questions regarding advanced grammar and vocabulary; these were taken from a resource book on language use (Fowler and Coe, 1980). All the fifteen students in group A scored between 31 - 40 correct answers, whereas the fifteen students in group B scored between 45 - 50 correct answers.

Source text

The source text was taken from a textbook on vocabulary strategies (Keen, 1985). The 570-word text on advertising, a popular topic among Spanish students, was divided into seven paragraphs, and its level of readability could be defined as intermediate, as a number of University lecturers pointed out (see Appendix A).

Procedure

Students read the text individually. After finishing, they had to write a summary, which was later handed in. Only the final version was observed, though some students had already created previous drafts. They had one hour to complete the task and were not allowed to use a dictionary.

As it was stated before, the main difference between both groups is the L2 proficiency level. However, a second striking difference is that, while group A had a set of instructions that were explained by the teacher before starting the task, group B did not receive any sort of help, fulfilling this task without having received any sort of previous specific summarising training.

As it is well-known summarising in L2 is a demanding task; because of this, we chose group A (the ones with a lower L2 proficiency level) to abridge the text following a set of instructions previously handed. The instructions were taken from the list suggested by Palmer (1996), although we added some modifications, trying to simplify the task as far as possible (see Appendix B).

Analysis of data

In order to analyse the thirty summaries collected, we followed these steps:

First of all, we analysed the quality of the summaries following Garner«s (1982) seminal article. We calculated the number of main ideas presented in each summary and then, we divided them by the total number of words used. Grammar mistakes and text elaboration were disregarded.
Secondly, we paid attention to the basic summarising strategies employed by the students. We followed the taxonomy stated by Palmer (1997), who classified them into three types: copy verbatim, generalisation and combination of two main ideas.
Thirdly, we observed how many subjects incorporated extra-textual information in their abridgements.
Finally, we examined if students stuck to the order of main ideas appearing in the source text or if they departed from it, by paying attention to the way they started and finished the summary as well as to its general development.

The general textual structure followed in the source text was a simple one. It begins with a general introduction which indicates that advertising is everywhere in America. In the following sections, the text author offers some pros (paragraphs 2 and 3) and cons (paragraphs 4-6) of advertising, whereas the final paragraph gives the general comment that commercials are an established part of modern life and concludes indicating that "sellers will always try to persuade buyers to purchase products. And critics (...) will always attempt to force business to keep their messages honest and clear" (Keen, 1985: 52).

Results

1. Quality of the summary

The source text was divided into seven paragraphs. After a thorough analysis, a number of University lecturers decided that this text had a total of seven main ideas which they considered should appear in an abridgement. In Table 1 we can observe how those seven main ideas have been reproduced in the students' summaries:

Table 1. Number of main ideas used in the summaries

                                    Group A   Group B
Total number of main ideas            85        99
Average of main ideas per summary     5.6       6.6
Total number of words                 1,228     1,406
Average of words per summary          81.8      93.7
Level of quality (Main ideas/words)   0.068     0.070

The data provide evidence that both groups of students understood the source text well enough to be able to summarise it, although their L2 proficiency level was quite different. In group A the average of main ideas per summary is 5.6 out of 7, presented in as few words as possible, whereas in group B, the average of main ideas per summary is 6.6 out of 7, though these texts were considerably longer (81.8 versus 93.7 words per abridgement, respectively). Nevertheless, if we pay attention to the different length of the summaries handed by these students, their level of quality is fairly similar. As Palmer (1997) commented, the length of these abridgement has a direct influence on the overall quality of the texts. Besides, as Garner (1982: 277) pointed out, the abridgements written by both groups of students could be defined as "middle-range efficiency summaries"; this means that most of these subjects were able to depict a high number of relevant ideas in a fairly moderate number of words.

Data also suggest that those subjects who received direct instructions on summarising performed almost as successfully as those people who, although having a higher level in English, had not received any direct training on how to abridge a text. These data seem to support the idea of the importance of teaching summarising strategies in the EFL classroom, in an attempt to make good all those problems observed among non-native summarisers when dealing with a text in a foreign language. Let us observe which are the strategies most often used by our subjects in order to abridge a source text.

2. Summarising strategies used by the students

In order to cut a long story short, we would like to point out some of the strategies used by students when facing a summarising task in a foreign language. Assuming the natural ability shown by younger students in order to omit irrelevant information, we have considered three different strategies carried out by our students when abridging a text: copy verbatim, combination of two main ideas, or generalisation of information in a single sentence. Table 2 displays the data observed in our analysis.

Table 2. Summarising strategies

                    Group A      Group B
Copy verbatim          7           73
Combination           17           18
Generalisation        56           23

The most important difference found out in this analysis was the high proportion of information copied verbatim observed among group B students, whereas their group A counterparts used generalisation in an equally high proportion. These results show that those students who received the set of instructions on how to draw up a good summary followed these rules and made the effort to use their own words, generalising information in order to create shorter, more concise texts, despite their lower L2 proficiency level. On the contrary, those subjects with an advanced L2 level just copied from the source text and condensed the information. One reason for the application of the copy verbatim strategy in group B may be found in the students' lack of knowledge of what summary writing actually entails. Nevertheless, we should also pay attention to a very usual device, fairly often related to faulty summarising ability, implied by the use of extra-textual information appearing within the abridgement

3. Extra-textual information

We define extra-textual information as those aspects of the general topic which did not appear in the source text, but have been included in the abridgement. In most cases they can show some reading comprehension problems, being often used as a device to get rid of difficult information that was not fully understood by the reader. Theoretically, our higher level students will be able to use it in fewer occasions throughout the experiment, using relevant information from the original source text.

Data show that just two students enrolled in group A furnished the summary with extra-textual information. In both cases we are dealing with text misinterpretations, something which confirms that all students knew that, when summarising, they did not have to incorporate additional information to their abridgements. Additionally they were also aware that they should not use their previous knowledge on the topic of the summary.

Even considering that the use of extra textual additions is a device used by students in order to increase the length of an abridgement--probably due to the lack of understanding of some of its passages--, we should also keep in mind that it can also show an interest for offering examples in order to complete relevant information. In those cases we can say that the use of extra-textual additions is proportionally inverse to the overall quality of the summary. Those students who tend to use extra-textual information delete basic data from the source text that should have appeared in their summaries. We should recommend our students to get rid of extra textual additions in their abridgements (Palmer, 1997). External information could only be accepted while clarifying aspects, and it would just increase considerably the length of the text, damaging the overall quality of the summary. If students are able to comprehend and select all the relevant information from the source text there is no need to use any kind of extra textual additions.

4. Rhetorical structure

How do students follow the organization of the text in their summaries? Do they stick to the source text order or do they depart from it? If we pay attention to Table 4, we observe that five students from group A did not comply with the source text order in their summaries. However, all those students with an advanced L2 proficiency level followed the layout observed in the source text; in all these cases, students started by drafting a general introduction, followed by an explanation of the different pros and cons of advertising (the main topic of the source text), and ending their texts with either a comment or a brief conclusion.

These results show that, although the natural tendency is to follow the source text organization, five subjects in group A departed from it and changed its general structure. In three cases subjects misunderstood the source text and, because of that, they changed the source text structure. The other two cases performed very well and, without missing many key ideas, they were capable of summarising the source text keeping the main ideas and changing its original structure.

Conclusions and Pedagogical Implications

The number of subjects in this study is so small that it is not possible to offer real generalisations or conclusions. However we can comment on the tendency we observed when we compared the summaries of the two groups.

The students' summaries under examination showed that both groups of students were able to get most main ideas from the source text in a moderate number of words, without including extra-textual information. Regarding the summarising strategies used by the students we observed that students in group A despite their lower L2 proficiency level, tried hard to use their own words when writing the summary, whereas those students in group B just copied from the source text. Finally we observed a natural tendency to follow the source text order in both groups, with the only exception of two students in group A who tried to depart from the source text order.

After analysing these data, we observed two aspects. Firstly, the L2 language proficiency significantly affected the summarising task; as data suggest, there was a tendency to perform quite well in group B (the advanced group) despite their lack of knowledge of what summary writing entails. Secondly we could also observe that the fact of having clear instructions regarding what is expected from a summary, helped group A (the ones with lower L2 proficiency) to enhance their abridging ability and perform almost as well as the students with advanced L2 proficiency.

Because of these results, we suggest that it may be relevant for our students to receive direct summarising instruction, being taught a series of instructions. In our opinion we can use this task as an activity to enhance both the reading and writing ability in the ESL classroom.

References

Amlund, J.T., Kardash, C.A.M., and Kulhavy, R.W. (1986) "Repetitive Reading and Recall of Expository Text". Reading Research Quarterly, 21 (1): 49-58.
Bracewell, R.J. (1981) "Writing as a Cognitive Activity". Visible Language, XIV: 400-422.
Brown, A.L., and Day, J.D. (1983). "Macrorules for Summarising Texts: the Development of Expertise". Journal of Verbal Learning and Verbal Behavior, 22: 1-14
Carey, L., and Fowler, L.J. (1986) "Strategies for Writing: Theories and Practices". College Composition and Communication, 37 (3): 302-314.
van Dijk, T.A. (1977a) Text and Context: Explorations in the Semantics and Pragmatics of Discourse. London: Longman.
van Dijk, T.A. (1977b) "Semantic Macro-Structures and Knowledge Frames in Discourse Comprehension". In M.A. Just, and P.A. Carpenter (eds.) Cognitive Processes in Comprehension. Hillsdale,NJ: Erlbaum.
van Dijk, T.A. (1979) "Relevance Assignment in Discourse Comprehension". Discourse Processes, 2: 113-126.
Donin, J., Bracewell, R.J., Frederiksen, C.H., and Dillinger, M. (1992) "Students' Strategies for Writing Instructions: Organizing Conceptual Information in Text". Written Communication, 9 (2): 209-236.
Eysenck, M.W., and Keane, M.T. (1990) Cognitive Psychology: a Student«s Handbook. Hillsdale: Erlbaum.
Flower, L.S., and Hayes, J.R. (1981) "A Cognitive Process Theory of Writing". College Composition and Communication, 32 (4): 365-387.
Fowler, W.S., and Coe, N. (1980). Test your English 3: Advanced. Edinburgh: Nelson.
Garner, R. (1982). "Efficient Text Summarization: Costs and Benefits". Journal of Educational Research, 75: 275-279.
Hayes, J.R., and Flower, L.S. (1983) "Uncovering Cognitive Processes in Writing: an Introduction to Protocol Analysis". In P. Mosenthal, L. Tamor, and S.A. Walmsley (eds.) Research on Writing: Principles and Methods. New York: Longman.
Hidi, S., and Anderson, V. (1986) "Producing Written Summaries: Task Demands, Cognitive Operations and Implications for Instruction". Review of Educational Research, 56: 473-493.
Keen, D. (1985) Developing Vocabulary Skills. Cambridge: Newsbury House Publishers.
Kintsch, W., and van Dijk, T.A (1978). "Towards a Model of Text Comprehension and Production". Psychological Review, 85: 363-394.
Palmer, J.C. (1996) The Use of Summarization Strategies in the Development of the Linguistic Competence of Business English Students. Colecci— Microfitxes. Castell—n: Universitat Jaume I.
Palmer, J.C. (1997) "Do I Have to Use my Own Words?- Summarising as a Tool to Enhance Writing Skills in the ESP Classroom". In R. Mart’n Chamorro and A. Romero Navarrete (eds.) Lenguas Aplicadas a las Ciencias y la Tecnolog’a: Aproximaciones, C‡ceres: Universidad de Extremadura.
Palmer, J.C. (forthcoming) "Cognitive Strategies in the Use of Summarising Techniques in the Business English Classroom.". Paper accepted for publication at the Proceedings of the VII International LSP Symposium. Copenhagen: CBS.
Sarig, G. (1988) Composing a Study-Summary: a Reading-Writing Encounter. Ramat Aviv: The Open University of Israel.
Stotesbury, H. (1990). Finnish History Students as 'Liminal' Summarisers on the Threshold of Academia. Jyvaskyla, Finland: University of Jyvaskyla. Reports from the Language Centre for Finnish Universities, No. 38.
Winograd, P. N. (1984). "Strategic Difficulties in Summarising Texts". Reading Research Quarterly, 19: 404-425.
Appendix A. Source text

Advertising
Anyone who lives in America is aware of advertising. Perusing newspapers and magazines, we see full-page ads urging us to buy clothes, autos, cigarettes, and kitchen appliances. Television and radio programs include commercials; we all have heard the phrases "bought to you by" and "sponsored by" hundreds of times. If we drive, we see road signs or billboards proclaiming the qualities of products or the location of restaurants or motels. If we commute on a transit system, we cannot help but notice the prominent signs displayed on the buses and subways. And in our mail, along with the bills and the letters, come shiny flyers and circulars promoting products and announcing sales.
The literal meaning of "advertise" is to make us turn toward something. When we see an ad, we turn our thoughts towards the message; we notice and remember what it says. At least, that is what the advertiser wants us to do. Sellers would have enormous problems transacting any business without advertisements. For example, if Apple or IBM or Texas Instruments introduces a new product like a personal computer, we would not know about it if these companies could not or did not advertise, so do sales.
Fortunately, the consumer benefits from advertising as well. Ads permit the public to buy intelligently. By reading the bank ads, for example, we might decide to transfer our money from our current bank to one offering better rates or more convenient hours. In addition, a traveller can save hundreds of dollars on transcontinental airfares by comparing the ads in the travel section of the newspaper.
Of course, nothing is perfect. Even the strongest proponents of advertising admit there are many problems. Some argue that commercials unnecessarily intrude into every walking minute of our lives. We simply cannot get away from the pounding, incessant messages. Because ads permeate radios and television, we find ourselves singing their silly jingles and repeating their "cute" lines. Sellers admonish us to buy through a profusion of techniques: hard sell, soft sell, music, comedy, and appeals to all our emotions and fears.
Some ads are potentially harmful. Perplexing or misleading sales pitches may lure unwary buyers into financial trouble. It is always best to remember: "caveat emptor" -let the buyer beware. Many commercials go far beyond the mere transmitting of information when they attempt to transform our values and attitudes. Cigarettes ads, for example, often imply that smoking is a manly or sexy habit. It is neither.
Because of these problems, some people have become extremely critical of commercials, especially those directed at children. As adults, we are often sceptical of what we read or hear in advertisements. Children, because they are not as mature or experienced as we are, cannot judge how reasonable or accurate ads are. If the man on TV says chocolate-covered-sugar-coated wheat toasts are healthful and nutritious, the children may very well believe it. Many parents feel sellers take unfair advantage of children's inability to evaluate what they see or hear.
Whatever problems commercials and ads may have, they are an established part of modern life and a permanent fixture in our daily life. Sellers will always try to persuade buyers to purchase products. And critics, continually arguing the pros and cons of advertising, will always attempt to force business to keep their messages honest and clear.

Appendix B. Set of Rules
List of rules that students should know in order to improve their summarising ability
1. Summarising means writing a shorter version of another person's work maintaining the gist of the information.
2. Summaries should not have repeated information.
3. We should start by finding the main topic of the summary.
4. Read the text thoroughly once in order to see what is the main topic. Read it again starting to underline all the important information.
5. To select important information use planning techniques, such as underlining or mapping.
6. Do not copy verbatim sentences from the original text. In case of doubt, paraphrasing is always better than copying.
7. Although the length of the summaries depends on the importance of the information appearing in the source texts, an average of 15-20 % of the total length of the source text will be advisable.
8. You should only use examples when it is absolutely necessary.
9. Avoid personal comments and opinions.
10. Maintain coherence and cohesion in your summaries.
11. Combining clauses can help you to shorten your summaries, but it is a difficult task, and has to be carried out with great care.
12. Only when you have understood the text completely you will be able to comprehend the different lexical, semantic, and grammar choices selected by the author. Once there, you will be able to choose your own decisions towards the creation of your very own summary.
These instructions have been adapted from Palmer (1996).

The Internet TESL Journal, Vol. IV, No. 1, January 1998
http://iteslj.org/
http://iteslj.org/Articles/Juan-TextSummary.html