Sociobiological bases of information structure

Abstract

This work faces the sociobiological bases of Information Structure. The term ―sociobiological‖ is here taken to encompass socio-interactional and processing implications of distributing sentence contents according to presupposition/assertion, topic/focus and given/new oppositions. The socio-interactional domain explores the interplay of Information Structure units and the linguistic encoding of evidentiality, whereas the processing perspective looks into the neurocognitive underpinnings that support the decoding of informational articulations in discourse. These two approaches are put together to set forth hypotheses on the emergence of Information Structure categories in human communication. The organization of the dissertation is summarized as follows. Chapter 1 provides an overview of the main theoretical literature on Information Structure from the earliest philosophical and Praguian traditions to the more up-to-date accounts. In this chapter, the correlation between information units and precise memory stores (expressly, Short-Term Memory and Long-Term Memory) is also laid out, together with the effects of information packaging on the storage and manipulation of information in the receiver‘s mental model of discourse. Chapter 2 outlines the interaction between the discourse realization of information units and the encoding of evidential meanings. On this purpose, a broad notion of evidentiality is taken into account (i.e. a notion embracing both the speaker‘s attitude towards a proposition and the grammatical marking of its source); precisely, a taxonomy of epistemic stances elaborated by Mushin (2001) is drawn upon. Two stances are contended to be crucial for the distribution of sentence contents into more or less relevant informational units; these are referred to as personal experience and factual stance. I suggest that the former correlates with the assertion and/or focalization of some information, whereas the latter more strongly relies on its presupposition and/or topicalisation. Chapter 3 presents experimental perspectives on the processing of Information Structure units, both from psycholinguistic and neurolinguistic perspectives. An introductory part gathers some the most far-reaching achievements reported in earlier and recent works on the subject. I put forth that these findings point towards two different trends in Information Structure processing. Using a terminology widely diffused within the purview of cognitive psychology and related disciplines, I called one 13 trend bottom-up, because it follows from processing operations that capitalize on the structural cues of sentence information; the other trend has been referred to as top-down because it reveals the influence of discourse-driven expectations on the processing of upcoming utterance contents. I attributed these two processing criteria to the experimental designs adopted in the reported studies, and, particularly, to the fact that in the bottom-up modality sentences were often processed in isolation, whereas in the top-down trend they were usually embedded in a wider context of discourse. In Chapter 4, two experiments are described that confirm the role of expectation-based parsing criteria when presupposed, asserted, topicalised and focalized information is processed. It is shown that, when information structures are compliant with the receiver‘s expectations on both activation state and information packaging of contents, sentence processing is easier as opposed to when they deflect from his pre-conceived mental representation of the discourse model. Chapter 5 represents the final part of the dissertation where the considerations developed in the foregoing chapters are built on to advance some possible evolutionary hypotheses of Information Structure in human communication. Here, the aforementioned socio-interactional (evidential) and processing-based arguments are recast as exogenous and endogenous forces or, in biological terms, as nurtural and natural biases, on the gradual shaping of Information Structure units. From the nurtural perspective, it is discussed that the presupposition/assertion and topic/focus dichotomies either emerged or have been exapted to modulate epistemic stances on communicated information. In this sense, in virtue of their discursive properties, topic and presupposition may have been selected (or exapted) to mark a pragmatic meaning of factual evidentiality, therefore reducing the speaker‘s commitment to the truth of a proposition. By contrast, focus and assertion may have been selected (or exapted) to mark a meaning of personal experience evidentiality, which increases the speaker‘s degree of commitment to truth. From a natural standpoint, the development of Information Structure is addressed against the background of processing constraints which I assume to be complied with by the above mentioned bottom-up and top-down processing modalities. In other words, there are conditions of sentence processing in which discourse-driven expectations cannot be relied on (as in the case of all-new sentences). In these cases, information structural cues may have appeared to serve the function of allocating processing 14 resources according to degrees of relevance of contents to the communicative tasks at hand. On this account, focus and assertion receive more attention because they are typically more purposeful in the current exchange; conversely, topic and presupposition require a lesser amount of attention, because they are less relevant to the communicative goal. In the top-down account, the units of Information Structure may have originated in order to facilitate the recognition of activation degrees of contents. More precisely, topic and presupposition allow relating some information to recently activated and previously shared contents respectively, whereas focus and assertion allow recognizing some content as new or unshared. Easing processes of mental recall of contents turned out to be particularly adaptive in a context of mainly oral communication (but the same can be said for written communication), which is generally more transient and ephemeral. So, the human attentional system exploits the cues provided by packaging in two ways: when previous expectations on the discourse model are not available, the processor allocates resources on the basis of information packaging instructions; on the contrary, if expectations on the activation state and packaging of contents can be properly computed, these guide processing. As for this latter condition, when expectations are not met, sentence processing is more demanding than it would be when information structure is consistent with the receiver‘s pre-conceived representation of the discourse model.