Facilita: reading assistance to the functionally illiterate
Willian Massami Watanabe · 2010 · Proceedings of the 2010 International Cross Disciplinary Conference on Web Accessibility (W4A) · doi:10.1145/1805986.1805997
Summary
This doctoral consortium paper presents Facilita, a web application that uses natural language processing (NLP) to automatically simplify Portuguese web content for functionally illiterate users. In Brazil, 21.7% of the population aged 15 and over is classified as functionally illiterate (defined by UNESCO as having fewer than 4 years of formal education), meaning they struggle to understand texts depending on size and complexity. While WCAG addresses readability through Guideline 14 (WCAG 1.0: "ensure documents are clear and simple") and Guideline 3.1 (WCAG 2.0: "make text content readable and understandable"), these require manual implementation by developers and content authors — a requirement that Brazilian accessibility surveys show is rarely met, and which becomes even harder in Web 2.0 contexts where non-expert users generate content. Facilita acts as an assistive technology layer that transforms inaccessible web text into more accessible text using four NLP operations: syntactic simplification (breaking complex sentences into simpler structures), automatic summarisation (reducing text length while preserving key information), lexical elaboration (providing definitions or simpler synonyms for difficult words), and named entity recognition (identifying and explaining references to people, places, and organisations). The tool is part of the larger PorSimples project on text simplification for Brazilian Portuguese.
Key findings
Two versions of Facilita were deployed: the core version implementing syntactic simplification and automatic summarisation, and an Educational version implementing lexical elaboration and named entity recognition. The paper is primarily a system description rather than an evaluation report, so detailed empirical results are not presented. The key technical contribution is demonstrating that NLP techniques can be applied as a client-side assistive technology to automatically adapt web content that was not authored with accessibility in mind — shifting the burden from content creators (who may not follow WCAG guidelines) to the reading tool. The approach addresses a limitation of WCAG's model: guidelines rely on voluntary compliance by developers and content authors, but surveys of Brazilian web development show accessibility awareness is very low and municipal government websites have poor accessibility.
Relevance
This work addresses one of the largest but least-discussed accessibility populations: people with low literacy. While most web accessibility research focuses on sensory and motor disabilities, functional illiteracy affects over a fifth of Brazil's population and hundreds of millions globally, creating barriers that are fundamentally about text complexity rather than interface mechanics. The approach of using NLP to automatically simplify content at the reader's end — rather than relying on content authors to write simply — is pragmatic and scalable. This anticipates the current interest in AI-powered content adaptation and text simplification tools. The paper also highlights that WCAG's readability guidelines (particularly the Level AAA Success Criterion 3.1.5 Reading Level) are among the hardest to implement because they require changing how people write, not just how pages are coded. Tools like Facilita offer a technological alternative when human compliance with plain language guidelines fails. The Portuguese language focus addresses an underserved linguistic community in accessibility technology.
Tags: literacy · cognitive accessibility · natural language processing · text simplification · readability · Portuguese · multilingual accessibility · digital inclusion
Standards referenced: WCAG 1.0 · WCAG 2.0