Ensuring cognitive accessibility in written texts is crucial for individuals with cognitive disabilities. The Easy-to-Read (E2R) Methodology provides guidelines for adapting texts, but its manual application is time-consuming and costly. This initial research explores the use of Large Language Models (LLMs) to infer and transfer rule-based linguistic knowledge representations (LKRs) from Spanish to German for two E2R guidelines: explanatory relative clauses and dialogue formatting. We evaluated three LLMs, finding that only Qwen2.5-7B-Instruct-1M generated a usable LKR, which still required extensive manual refinement. The results show that while the LLM was helpful in identifying structures, it struggled to produce grammatically correct adaptations. The refined rules performed better, but still needed human adjustment. A validation with two German E2R experts confirmed the potential of the proposal, while also pointing out common errors. Additionally, a user-based evaluation is designed and planned to assess whether the adapted texts improve readability for individuals with cognitive disabilities.