In this paper, we use survey data from the Okazaki Honorific Language Survey to clarify changes in the formalization of linguistic behavior when explaining situations in response to destination questions. For the survey data on the "going to city hall" situation, we analyzed the changes in the response utterances by dividing them into the components of [starting a conversation] and the [person being asked], [content of the errand], and [destination] of the situation explanation. As a result, it was revealed that the elements of [person being asked] and [content of the errand] gradually changed to not being mentioned, and the element of [destination] changed to clearly stating "going to city hall". From the perspective of the formalization of greeting expressions, this change can be said to indicate a formalization pattern in which peripheral elements related to going out are no longer mentioned, and only the [destination], which is the core of the answer to the destination question, is specifically stated. Through the verification of this paper, we have demonstrated the possibility of language change research, which can gain a research perspective from the results of surveys of changes over time on language phenomena that are difficult to understand through literature surveys alone.

View full abstract

In this study, we performed the task of simplifying standard news broadcasts by NHK into plain Japanese using ChatGPT. Specifically, we quantitatively analyzed four types of news articles: 1) standard news, 2) news simplified by the GPT-4 model, 3) news simplified by the GPT-4o model, and 4) 'Easy Japanese News' crafted by humans, based on the readability analysis results. The analysis revealed that the articles generated by GPT-4 were the most readable, while the standard news was the least readable. Furthermore, the text generated by GPT-4o was closer to the 'Easy Japanese' created by humans. Additional analysis of textual features indicated that GPT-4o simplified the text by adjusting factors such as average sentence length, while faithfully reflecting the lexical attributes of the standard news. These results suggest that generative AI, through large language models, is not only improving in surface task performance but also evolving towards human-like reasoning and language generation capabilities.

View full abstract