In an increasingly networked world, people who cannot participate in written communication are particularly at risk of social exclusion. Like other concepts of barrier-free communication, Easy Language (“Leichte Sprache”) attempts to enhance perceptibility and comprehensibility of texts for people with reading impairments by means of a rule-based reduction of complexity on the text side. The underlying assumption of simplified languages is that the reduction of linguistic complexity correspondingly reduces cognitive processing costs. However, this interplay between linguistic complexity and cognitive processing costs still needs to be investigated by empirical research as up to date there are only a few studies investigating the perception and reception of Easy Language, mostly using offline methods (e.g., questionnaires or retrospective interviews). In contrast to offline methods, which are only capable of assessing comprehension products, online methods allow researchers to track what a participant is focusing their attention on at any given time and to thereby develop a detailed representation of the time-course of cognitive language processing. In our paper, we aim to point out how different online methods (eye-tracking, EEG and fMRI) can be used for investigating the empirical validity of the postulated rules for Easy Language by evaluating cognitive processing efficiency. Besides discussing the applicability of those neuroscientific online methods in Easy Language research, we discuss the importance of collecting personal and neuropsychological data to gain detailed profiles of the participants and therefore not only contribute to the explanation of variance but furthermore to determine the role of neuropsychological skills on reading proficiency. For each online method we elaborate basic principles, discuss some of the main findings in cognitive sciences and demonstrate the greatest advantages but also restrictions of the method and challenges related to the data collection process with impaired participants. Furthermore, we outline current challenges in Easy Language research and summarize remaining research gaps. On a final note, we emphasize that it is both the establishment of online methods and the data triangulation in Easy Language research that enable researchers to gain a profound insight into the cognitive processing of simplified languages.
Two experiments were conducted to investigate the effects of visual segmentation, complexity, and context on the cognitive processing of compounds in Easy Language. By presenting compounds in different boundary conditions, we determined whether a segmentation cue facilitates the processing of compounds presented with and without contextual information. The study was conducted with unimpaired adults and with hearing-impaired pupils, representing one of the Easy Language target groups. The results indicate that visual segmentation facilitates processing of compounds for pupils with low literacy skills. However, they only benefit from segmentation when morpheme boundaries are marked in a subtle way, i. e., without strikingly deviating from the standard version. Pupils with higher literacy skills and unimpaired adults do not profit from segmentation. Even though hyphenation slows down compound processing for unimpaired readers, initial processing advantages of hyphenated over concatenated compounds emerged, which is explained by the fact that hyphenation forces a morpheme-based access and enables fast recognition of the compound’s first constituent. However, it hinders readers from accessing the compound via the direct route and thus slows down the processing of the compound as a whole.
1 January 2025 in SKASE Journal of Translation and Interpretation
In this paper, we present the results of a study evaluating intralingual machine translations of health information texts into Plain German. In our study, we compare machine-translated simplified texts with those simplified manually by human translators, as well as with the original, unsimplified texts. We compare the output of four different machine translations systems and assess the translation quality using various criteria, including translation errors, readability, and syntactic complexity. The study reveals that from the four analysed machine translation systems, ChatGPT performed worst. Our results also suggest that fine-tuning a model with task-specific and domain-specific data improves the translation quality.