In this paper, we investigate the factors that influence interactants’ choice of syntactic variants when using a predicative adjective construction like that is okay. In German colloquial conversation, such constructions can occur as full sentences with subject, copula and adjective (das ist gut ‘that is good’); with topic drop consisting of copula and adjective (ist gut); or as fragments consisting only of the adjective (gut). We present findings from a corpus of colloquial speech between fellow students showing that the interactional function of listener feedback has a higher predictive power in accounting for the use of fragments vs. fuller structures than adjective semantics (descriptive vs. evaluative), propositional structure (reference to individual or propositionally structured referent), and predictability in terms of adjective frequency. Moreover, we find that fragments consisting of evaluative adjectives show a clear tendency to be grounded in the here-and-now of the current situation, whereas fuller structures are more apt to express evaluations grounded in past experience. We argue that fragments are formally optimized to convey expressive actions such as listener feedback and other ad-hoc evaluations.

In this paper, we compare the results of three studies. The first explored feature-conditioned distributions of referring expression (RE) forms in the original corpus from which the contexts were taken. The second is a crowdsourcing study in which we asked participants to express entities within a pre-existing context, given fully specified referents. The third study replicates the crowdsourcing experiment using Large Language Models (LLMs). We evaluate how well the corpus itself can model the variation found when multiple informants (either human participants or LLMs) choose REs in the same contexts. We measure the similarity of the conditional distributions of form categories using the Jensen-Shannon Divergence metric and Description Length metric. We find that the experimental methodology introduces substantial noise, but by taking this noise into account, we can model the variation captured from the corpus and RE form choices made during experiments. Furthermore, we compared the three conditional distributions over the corpus, the human experimental results, and the GPT models. Against our expectations, the divergence is greatest between the corpus and the GPT model.

The generation of referring expressions (REs) is a non-deterministic task. However, the algorithms for the generation of REs are standardly evaluated against corpora of written texts which include only one RE per each reference. Our goal in this work is firstly to reproduce one of the few studies taking the distributional nature of the RE generation into account. We add to this work, by introducing a method for exploring variation in human RE choice on the basis of longitudinal corpora - substantial corpora with a single human judgement (in the process of composition) per RE. We focus on the prediction of RE types, proper name, description and pronoun. We compare evaluations made against distributions over these types with evaluations made against parallel human judgements. Our results show agreement in the evaluation of learning algorithms against distributions constructed from parallel human evaluations and from longitudinal data.