Functions of Language - Current Issue

This paper offers a conceptual clarification and an operationalization of a classic functionalist generalization: the one-new-idea constraint (Chafe 1982, 1987, 1994). While often referenced, this constraint has not, to date, been sufficiently spelled-out or made sufficiently precise to become testable. This paper aims to fill this gap by, first, differentiating the concept of idea into discourse-level and lexical-semantic ideahood. Secondly, we propose that the one-new-idea constraint (onicon) should be investigated in Chafe’s (1994) original encompassing sense, where not only referents, but also states and events are tracked. The third clarification applies to the unit of application: As a constraint on speech, this should be the intonation unit as suggested by Chafe, rather than the clause or sentence as suggested elsewhere (e.g. Givón 1990). Based on these conceptual refinements, we develop an operationalization for testing the onicon: the tonic (Testing the One-New-Idea Constraint) annotation scheme. This scheme not only allows for in-depth investigation into the robustness of the onicon, but also into information flow on both discourse as well as lexical-semantic levels more broadly. tonic thus serves as a new tool for more holistic research into information flow.

The study explores Chafe’s one-new-idea constraint in two short extracts of natural multimodal interaction in two Trans-Himalayan (Tibeto-Burman) languages: Burmese and Anal Naga. It corroborates the overall tendency in both extracts for one new idea per intonation unit on average. However, the examination of frequent exceptions to this tendency indicates that this is not a cognitive constraint, as informatively much denser intonation units are processed with no challenge for interactants. The counterexamples suggest that the phenomenon emerges from more basic factors, and primarily from attention allocation, joint attention alignment and its interactional monitoring, and intonational partitioning for turn-management. As such, the observed tendency is a joint product of communicative and cognitive factors that shape conversation structure into sequences of locally monitored attention shifts.

The One-New-Idea-Constraint (ONICON) is proposed to be an absolute universal on human cognitive capabilities. While the ONICON predicts that we maximally realise one new idea at a time, it does not account for the maximal amount of information that can be realised at any point in discourse. To help us better understand factors influencing information flow in discourse, we may instead wish to turn our focus to verbalisation, the cognitive process behind this act. Examining verbalisation more closely can in turn allow us to better assess the extent to which the ONICON can hold. Using corpus data from three Trans-Himalayan languages, I explore how we choose to verbalise particularly information-rich intonation units and what factors may influence this process. Through investigating how information-rich intonation units are verbalised, I highlight ways in which the ONICON may be refined or challenged to better capture everyday language use.

It is generally agreed that intonation units tend to correspond to processing units (Chafe 1979, 1987, 1994, 2018; Inbar et al. 2023; and many others). As noted by other authors in this collection, the one new idea at a time pattern shows up robustly across typologically quite different languages. It is seen, for example, in languages with complex nominal expressions (Louagie & Reinöhl 2022) and those with serial verb constructions (Pawley 1987, 2009; Givón 1991; Reinöhl et al. this volume). Similar prosodic packaging can also be seen in languages with very elaborate morphologies. It can, furthermore, account for the formation of some of their typological structure. Such effects are illustrated here with conversational data from Mohawk, a polysynthetic language indigenous to northeastern North America.

The one-new-idea constraint (ONICON) is a hypothesis about how much information we can package in one intonation unit (Chafe 1982, 1992, 1994). Chafe (1994) suggests that constructions containing multiple new content words do not necessarily violate the ONICON. Examples include prefabs as well as constructions consisting mostly of low-content elements, and no more than one high-content word (as in, e.g. light verb constructions). We argue in this paper that there is another construction type — functor–argument metaphors (Reinöhl & Ellison 2024) — which does not violate the ONICON, even while they can also consist of more than one new content word. In metaphorical functor–argument constructions, such as to arrive at a conclusion, the functor arrive loses semantic features — in this case, movement in space — through the clash of its literal sense with a non-spatial ‘goal’. The resultant lowered content of the functor renders functor–argument constructions like ones containing maximally one high-content lexeme to start with. We will show that functor–argument metaphors also share certain properties with prefabs. We conclude that functor–argument metaphors need to be added as a third pseudo-exception type, leading us to an enriched understanding of the ONICON, ideahood, and holistic processing of multi-word construction types in speech.

We investigate a notion related to Chafe’s One New Idea Constraint (ONICON), namely that clauses only contain one new idea at a time. We investigate both the nominal and the verbal domain in discourse production data from 16 diverse languages. We find no convincing evidence for a hard constraint on the number of new ideas per clause, nor for speakers actively managing information flow in discourse. The absence of an underlying ONICON is suggested by (1) the fact that NPs co-occurring in a single clause do not have more accessible referents, (2) that observed distributions of information can likewise be reproduced by a randomised mechanism, and (3) the absence of a presentational referent activation pattern. A more likely account of observed discourse production pattern is producers’ overall goal to be both contentful and coherent, the former implying the mentioning of a considerable number of entities and the latter implying predications about these entities. It is in this sense that there is audience design, rather than in an avoidance of information overcharge.