This extended abstract is a contribution to the Easy-to-Read on the Web Symposium. The contents of this paper were not developed by the W3C Web Accessibility Initiative (WAI) and do not necessarily represent the consensus view of its membership. For more information about RDWG symposia and publications, see the RDWG Frequently Asked Questions (FAQ).
The technical aspects of Web content accessibility are discussed since many years and are addressed by international guidelines and legal regulations in many countries. The importance of understandable content and accessible information for persons with learning difficulties has only recently begun to receive increased attention. The rules and guidelines for understandable Web content are more heterogeneous. Often they were defined as ad-hoc rules and lack scientific evidence.
More specific guidelines can help address the increased demand for Easy-to-Read (E2R) information on the Web. They are helpful in teaching people to write E2R. They also simplify the checking of E2R texts, and can be used to implement tools to support less experienced authors (Nietzio et al., 2012).
This paper analyzes the differences between E2R and Plain Language (PL) with regard to target groups and guidelines. We present a linguistic analysis of selected criteria to get a better understanding of the guidelines for the two language levels.
Over the last decades a more flexible language provision evolved, considering the needs of people with permanently or temporarily limited reading proficiency. Underlying guidelines were often developed heuristically and address especially print documents. Only in recent years E2R and PL started to focus other media (Inclusion Europe 2009), (Matausch & Peböck, 2007), (Tronbacke et al., 2010).
Whereas E2R and PL movements around the world are common (Plain Writing Association), PL is still underrepresented in the German speaking area and needs further development (Capito 2012). A central aspect of the German approach to E2R is the focus on proofreading. All text should be checked by readers from the target group.
People with intellectual and learning disabilities are the main E2R target group. Additionally, pre-lingually deaf persons, deafblind persons, persons with dementia belong to the group (Tronbacke et al., 2010). Individual cognitive abilities such as attention span and memory also play a role.
Practice shows that non-alphabetized people and immigrants with a different native language benefit of E2R. In 2011 a German study found that approximately 15% of the population are functionally illiterate Grotlüschen & Riekemann, 2011). These people have poor literacy skills and make little or no use of their reading proficiency. E2R and PL are crucial to involve them in society.
A person with less reading experience is more likely to belong to E2R or PL target group. Practice demonstrates that the more an individual is introduced to textual information, the more these persons are motivated to read by themselves and might move from the E2R to the PL target group.
This section presents the guidelines that were selected for the linguistic analysis in our approach with a particular focus on the similarities and differences of the guidelines for E2R and PL. The guidelines differ to some extent, for instance concerning font-size and length of lines. In contrast to E2R, PL information does not make use of line-breaks after a unit of meaning.
Short sentences are a common criterion for E2R and PL. E2R material usually presents one idea per sentence. PL information does the same as this makes the text easier to follow. Nevertheless, if the verification with the PL target group approves a given sentence, the usage of this rule will play a minor role.
The use of one dependent clause is allowed in E2R and PL. In E2R material dependent clauses often are built with "if-then"-combinations to emphasize the chronological sequence, which is one of the most important principles of E2R. Dependent clauses starting with "although", "nevertheless", "since", and "in order to" is discouraged in E2R. This is contrary to PL where these conjunctions can be used.
E2R and PL guidelines require active language, because this activates the reader. Practice supports this assumption. Nevertheless also indirect speech and passive voice is understood by the target groups depending on the context.
Our goal is to refine the guidelines for E2R and PL with special focus on the differentiation of the two language levels. This can contribute to the production of texts that are better suited for the target audience.
Our research was carried out on a parallel corpus of texts about the Austrian disability law, which is available in E2R and PL. We assume that by analyzing the two types of text we can find empirical evidence for the implicit knowledge of authors and proofreaders that can be used to develop more detailed rules, especially in areas where the underlying rules are similar:
These rules do not strictly ban the use of dependent clauses or passive voice. The decision is left to the expertise of the authors and the persons from the target group who check the text.
First we analyze the texts to identify the similarities and differences with respect to use of dependent clauses and passive voice. Then we describe the findings as potential rules and compare them to the actual guidelines applied in the texts.
The number of ideas in a sentence can be measured by the number of (dependent) clauses in the sentence (Nietzio et al., 2012). In both the E2R and the PL material we found that about 70% of the sentences consist only of a main clause. The E2R material has 25% sentences with one dependent clause and about 5% with two dependent clauses. Only very few sentences are more complex, so that we assume they passed the quality assurance unnoticed. In the PL material the sentences are more complex. 20% have one dependent clause and there are 5% with two and another 5% with three or more dependent clauses. This shows that the same underlying rule leads to different texts because it is applied less rigorous to PL material.
Conditional clauses occur frequently in both types of text. In the E2R material one third of the conditional sentences uses explicit "if-then"-construction to emphasize the conditional relation. In the PL material less than 10% use this construction. Instead there is more variety in the expression of the conditional relationship ("in case", "only if" or use of word order to express condition) accounting for 15% of the cases. Again we can see that the same rules are applied but the PL version is less strict.
About 15% of the sentences in the PL material use passive voice, but only about 5% in the E2R version, leading to the conclusion that avoiding passive contrutions is less important in PL. Almost half of the sentences in passive voice (in both language levels) contain an explicit reference to the agent ("by") or the situation ("in", "at") to support the understanding of the sentence. For instance: "The assisted transport is organized by the state." This approach is not described in the guidelines yet and might be useful to add as new rule.
The detailed analysis also discovered a small number of deviations from the guidelines that were missed in earlier proofreading.
In this paper we suggested a method and applied it successfully to a small set of texts and rules. In the future the approach can be extended to provide empirical justification for a wider selection of E2R and PL rules. It can also be used to refine existing guidelines and elaborate on the distinction of language levels.
It is not always simple to find parallel texts in E2R and PL for comparative analysis. A broader basis of text would be helpful to validate our approach and the findings.
Another limitation is that our results are very specific for the German language. The more specific the E2R rules are, the less they can be generalized to other languages. "Use short sentences" can be applied in many languages but rules that relate to grammatical concepts such as type of conjunction or syntactic construction are language-specific. However, the approach that was used should be suitable for application to other languages.
Our ideas for future research go beyond the analysis of written text. In addition to the linguistic features of E2R, also the presentational features such as line-breaks to structure sentences or boldface for emphasize should be explored. The accuracy of E2R and PL criteria for digital information should be reviewed, especially regarding the coverage of E2R in WCAG. Moreover, it would be interesting to investigate how the reading level defined in WCAG 2.0 relates to the criteria for PL. Finally, the use of tools to assist authors and support the linguistic analysis could be explored further.
This work is supported by the Department of Social Affairs of the Upper Austrian Government and by the project Di-Ji ("Digitally informed — integrated at work") funded by the German Federal Ministry of Labour and Social Affairs.