The present paper focuses on the study of opinion dynamics and opinion shifts in social media in the context of climate change discourse in terms of the quantitative NLP analysis, supported by a linguistic outlook. The research draws on two comparable collections of climate-related social media data from different time periods, each based on trending climate-related hashtags and annotated for relevant sentiment values. The quantitative computer-based research methodology has been supported by a language-based perspective in the pragma-linguistic form. The research shows that the latter data source, for the majority of identified topics, exhibits a significant reduction in negative sentiment and a dominance of positive sentiment, i.e., a potential temporal evolution in public sentiment toward climate change. To achieve this, we used a BERT-based clustering approach to identify dominant themes within a combined dataset of tweets from both periods. Subsequently, a unified sentiment classification framework using a Large Language Model (LLM) was applied to reclassify all tweets, ensuring consistent and climate-specific sentiment analysis across both datasets. This methodology allowed for a coherent comparison of public attitudes and their evolution in different time periods and thematic structures.

The paper focuses on testing the use of conversational Large Language Models (LLMs), in particular chatGPT and Google models, instructed to assume the role of linguistics experts to produce opinionated texts, which are defined as subjective statements about animates, things, events or properties, in contrast to knowledge/evidence-based objective factual statements. The taxonomy differentiates between Explicit (Direct or Indirect), and Implicit opinionated texts, further distinguishing between positive and negative, ambiguous, or balanced opinions. Examples of opinionated texts and instances of explicit opinion-marking discourse markers (words and phrases) we identified, as well as instances of opinion-marking mental verbs, evaluative and emotion phraseology, and expressive lexis, were provided in a series of prompts. The model demonstrated accurate identification of Direct and Indirect Explicit opinionated utterances, successfully classifying them according to language-specific properties, while less effective performance was observed for prompts requesting illustrations for Implicitly opinionated texts.To tackle this obstacle, the Chain-of-Thoughts methodology was used. Requested to convert the erroneously recognized opinion instances into factual knowledge sentences, LLMs effectively transformed texts containing explicit markers of opinion. However, the ability to transform Explicit Indirect, and Implicit opinionated texts into factual statements is lacking. This finding is interesting as, while the LLM is supposed to give a linguistic statement with factual information, it might be unaware of implicit opinionated content. Our experiment with the LLMs presents novel prospects for the field of linguistics.