Text Modification in Foreign Language Teaching

Author: Rocio Ortega de Toro


The key points to the paper: introduction: research in progress; the importance of reading comprehension; authentic vs. modified: previous studies; criteria for modifications; preparations for the first experiment; conclusions so far.

This article was added to our website on 20/12/02 at which time all links were checked. However, we cannot guarantee that the links are still valid.

Table of contents

This paper was originally presented at the Setting the Agenda: Languages, Linguistics and Area Studies in Higher Education conference, 24-26 June 2002.


This paper is concerned with the issue of modification in texts that are normally used for foreign language teaching classroom. The inspiration for this topic emerges from my experience as a foreign language assistant in Spanish at The University of Sheffield. My daily practice with Post A level students raised some issues on the subject of reading comprehension problems, particularly, how using authentic texts was affecting my students' ability to understand enough content to move forward in the lesson.

The main difficulty I have experienced as a language tutor was that when students do not understand a text it becomes impossible to move forward with the plan of activities based on that specific text. A wider consequence which is that these 'out of reach' texts create a negative experience with reading in the target language for the student. Most importantly the ultimate effect is the lack of progress in the process of acquisition in the language.

The importance of reading comprehension

As Adel Tweissi states: 'it is well established in the literature on second language acquisition that comprehension of a message by the language learner is an essential condition on the acquisition process' (Tweissi, 1998).

If we accept this statement as true, then we agree on the importance of reading comprehension. Most of the time, reading comprehension is at the very heart of the acquisition process and it is through reading that a vast amount of that process takes place.

Text selection and its problems

Choosing a text on which to develop a fifty-minute class has never been easy. I have often found myself using what I believed to be excellent texts for oral discussion and subsequent activities in the class, only to later realise the complexity that they presented to students.

One such example is the 'kleenex text' - a text selected from a Spanish newspaper about the life of a couple who works selling tissues at traffic lights in Spain, something quite common in this country and which I thought would raise interesting discussions. The text had 111 words, 26 of them were not understood by over half of the students, 19 were not understood by almost all of the students. 10 out of the 19 were slang words. This is an example of text that should not be used in its authentic version by students of this level because it hindered comprehension and subsequently impeded moving forward in the class.

Authentic vs modified texts

A significant amount of research has been carried out on the question of the use of authentic versus modified texts. A number of studies e.g. Coleman (1962) and Ali (1994) show that lexically and/or syntactically modified texts do improve reading comprehension. On the other hand studies like the one carried out by Blau (1982) show that the most simplified version of a text is not necessarily the one that obtains better comprehension results by students reading it.

A more recent study by Adel Tweissi (1998) confirms these results. Authentic texts are found to be not always the most appropriate, especially for students not linguistically mature enough to face them. However, amongst the different levels of simplifications/modifications in the various versions, the simplest was not the one that produced the best results in terms of reading comprehension. It was with one of the modified versions that students scored highest regarding comprehension. In all cases, the students reading the authentic version had the worst results.

All these results come to show that there is something about the type and amount of modification that should be looked into and analysed carefully. These are the factors that might hold the clue as to how we could modify a text to suit the students' level and ultimately allow for progression.

Types of modification: criteria for modification

It is the aim of this study to test the effect of certain types of modification on students' reading comprehension. In order to identify the types of modification that are beneficial for language acquisition, we first need to identify the criteria on which to base our modifications. In this respect we decided that there is one main factor to take into account: the positive and negative relationships between L1 and L2. This is an idea that emerged in some very valuable discussions that I have maintained with Dr Anthony Bruton (University of Seville) on this topic.

In this case L1 is English and L2 is Spanish but the same concept could be applied to other pairs of languages.

What do we mean by positive and negative relationships?

Here is where it is important to consider which two languages we are dealing with. The L1 does inevitably have an influence on the whole process of acquisition of a given L2. This is why considering the linguistic relationship between both languages is crucial. There are aspects of the students' L1 that will certainly have an influence while they are learning an L2. It normally happens when students try to apply established patterns of their mother tongue to the new language being learnt. This is normally done at syntactic and/or lexical level of the language. The concept of interaction between L1 and L2 is from our point of view applicable not only to the example of Spanish and English but to other pairs of languages.

Besides this main criterion there are other factors that should be taken into account when introducing modifications. An important one is to be sure that we achieve a balance between unfamiliar and familiar elements in a given text. One can even play around with unfamiliar elements and modify them in some cases and not in others. For example, in our first experiment there was a word that appeared frequently throughout the text. We decided to modify it towards the first part of the text to avoid significant misunderstandings from the very beginning, but then we left the unmodified word to introduce a new element in their vocabulary.

Some examples of types of words that can produce a positive or negative relation between the L1 and L2:

false friends negative
cognates positive
colloquialisms negative

Test 1

Pre-test issues

The first experiemnt was prepared following Hughes's guidelines from his book Testing for Language Teachers (1989).The factors taken into consideration were:

  • operations: multiple choice, short answer, guided short answer;
  • type of text: journalistic because is readily available and they present up to date issues;
  • addressees: First year post A level Hispanic studies students;
  • topics: The use of Internet at work (no background knowledge from my classes because we had not discussed this topic at all before).

Among the many variables to be taken into consideration with regard to the reliability of results were:

  • Background knowledge effect on their answers to the reading comprehension questions. To avoid this effect it was important that we used a text that was completely new to them in the test.
  • Type of exercise used: the actual exercises or questions in the tests should not be a barrier for the student to answer correctly. This is why we used the types of exercises detailed above.
  • Use of context to guess meaning. In order to keep this variable under control the test had not only general reading comprehension questions but also specific vocabulary questions. This would allow for differentiating between when the student knew the answer to a general reading comprehension question thanks by the context or as a result of actually knowing meaning of a given word.

Analysis of first experiment

The test was administered to 48 first year post A Level students in the Department of Hispanic Studies in The University of Sheffield. This number of students was divided into two groups. One group was given the authentic version of the text: this was the control group. The second group was given the modified version of the text: this was the research group.

Control group (authentic version of text)

In the results, we were looking for the relationship between correctness in the answers to specific items of vocabulary and in answers to the general meaning questions. Approximately 50% of students that did have some problems with vocabulary items did indeed answer incorrectly one or more of the general meaning questions. These results indicate that almost half of the students' comprehension was significantly hindered when they dealt with the authentic version of the text. The fact that the same thing did not affect the comprehension of the other half of the students is also very significant. In my opinion this happened due to the greater competence that those students have of the L2. These students had problems with specific items of vocabulary too but that did not affect their general understanding of the text. This draws us to the presumption that the lower the student's level of language, the more useful the modifications will be.

Research group (modified version of text)

Results from the students dealing with the modified version of the text show various interesting things:

  • The vast majority of students did not have any problem with specific items of vocabulary, presumably because they had been modified.
  • Consequently the general understanding of the key ideas in the text was not affected. The students scored indeed very highly on the reading comprehension questions.
  • One of the modifications introduced did not have the effect that it should have had, i.e. improving reading comprehension. Instead, that modification caused bad answers in the students because it contained a word that can mean different things in Spanish (terminos del contrato). This is something to learn from and it shows how carefully we must work on the modifications.

Conclusions so far

The above results lead us to the conclusion that modifications do improve reading comprehension and in the case of the oral classes it would have a positive effect. However, we remain uncertain as to how much the students will learn from modified versions. In my opinion the modified version of my experiment was only slightly different from the authentic version. This means that the student is being exposed to a more or less 'authentic looking text' even in the modified version. There are some options that can be combined with the use of modified texts, for example, teachers can show students the modifications that have been carried out in a text as part of the lesson methodology. In this way, they would definitely also be exposed to new items. Another option is the one I have already mentioned, i.e. introducing modifications at one stage of the text and then introduce the authentic equivalents later in the text.

Perhaps the most important conclusion so far is that the students that will benefit the most from these types of modifications are those that are of under post A level standard. The same test that we have just done will be administered to second-year post beginner students in the same department later this year and then we will be able to compare and analyse the difference. This is not to say that other students at other levels cannot also benefit from the modifications, but the way these are introduced must be carefully thought of.