Miscommunication phenomena such as repair in dialogue are important indicators of the quality of communication. Automatic detection is therefore a key step toward tools that can characterize communication quality and thus help in applications from call center management to mental health monitoring. However, most existing computational linguistic approaches to these phenomena are unsuitable for general use in this way, and particularly for analyzing human–human dialogue: Although models of other-repair are common in human-computer dialogue systems, they tend to focus on specific (...) phenomena, missing the range of repair and repair initiation forms used by humans; and while self-repair models for speech recognition and understanding are advanced, they tend to focus on removal of “disfluent” material important for full understanding of the discourse contribution, and/or rely on domain-specific knowledge. We explain the requirements for more satisfactory models, including incrementality of processing and robustness to sparsity. We then describe models for self- and other-repair detection that meet these requirements and investigate how they perform on datasets from a range of dialogue genres and domains, with promising results. (shrink)

A key requirement for the automatic generation of argumentative or explanatory text is to present the constituent propositions in an order that readers will find coherent and natural, to increase the likelihood that they will understand and accept the author’s claims. Natural language generation systems have standardly employed a repertoire of coherence relations such as those defined by Mann and Thompson’s Rhetorical Structure Theory. This paper models the generation of persuasive monologue as the outcome of an “inner dialogue”, where the (...) author attempts to anticipate potential challenges or clarification requests. It is argued that certain RST relations such as Motivate, Evidence and Concession can be seen to emerge from various pre-empting strategies. (shrink)

Feedback such as backchannels and clarification requests often occurs subsententially, demonstrating the incremental nature of grounding in dialogue. However, although such feedback can occur at any point within an utterance, it typically does not do so, tending to occur at Feedback Relevance Spaces. We present a corpus study of acknowledgements and clarification requests in British English, and describe how our low-level, semantic processing model in Dynamic Syntax accounts for this feedback. The model trivially accounts for the 85% of cases where (...) feedback occurs at FRSs, but we also describe how it can be integrated or interpreted at non-FRSs using the predictive, incremental and interactive nature of the formalism. This model shows how feedback serves to continually realign processing contexts and thus manage the characteristic divergence and convergence that is key to moving dialogue forward. (shrink)

Quotation has been much studied in philosophy. Given that quotation allows one to diagonalize out of any grammar, there have been comparatively few attempts within the linguistic literature to develop an account within a formal linguistic theory. Nonetheless, given the ubiquity of quotation in natural language, linguists need to explicate the formal mechanisms it employs. The central claim of this paper is that once one assumes a dialogical perspective on language such as provided by the KoS (KoS is not an (...) acronym, despite emphasizing a Konversationally Oriented Semantics) framework, formalized in a rich type theory like Type Theory with Records, much of the mystery evaporates. In particular, one can utilize as denotations for quotative constructions entities that are independently motivated for dialogue processing—utterance types and locutionary propositions, Austinian propositions about speech events. (shrink)

Clarification requests, queries posed in response to a “problematic” (misheard, misunderstood, etc.) utterance, are a challenge to mainstream semantic theories because they call into question notions such as “shared content” or “the context.” Given their strong parallelism requirements, elliptical clarification requests introduce in addition significant complexities concerning the need for long-term maintenance of non-semantic information in context. In this paper, we consider a puzzle concerning the emergence of elliptical clarification requests in child English: Data from the Belfast and Manchester corpora (...) from CHILDES demonstrate that reprise fragments, the highest frequency clarification request construction among adults, emerges with significant delay in comparison with reprise sluices, bare wh-phrases used to request clarification. This is a puzzling finding a priori: first, since reprise fragments are by most plausible measures the most readily available clarification request form in that it involves mere repetition of material primed by the previous utterance and children have the ability to repeat parts of the previous utterance from the earliest stages of speech. Moreover, as we show, reprise fragments predominate—in some cases vastly so, over sluicing in the input of fragmentary clarification requests available to the child—and it would, therefore, be difficult to construct an explanation for order of emergence primarily based on frequency of the construction in the children's input. Our account is based on grammatical explication of the difference between the constructions, how constructional difference is represented in a hierarchy of constructions, and in terms of a notion of semantic complexity that suggests reprise sluicing is, in fact, less complex than reprise fragments. (shrink)

In a series of papers, Donald Davidson :3–17, 1984, The philosophical grounds of rationality, 1986, Midwest Stud Philos 16:1–12, 1991) developed a powerful argument against the claim that linguistic conventions provide any explanatory purchase on an account of linguistic meaning and communication. This argument, as I shall develop it, turns on cases of what I call lexical innovation: cases in which a speaker uses a sentence containing a novel expression-meaning pair, but nevertheless successfully communicates her intended meaning to her audience. (...) I will argue that cases of lexical innovation motivate a dynamic conception of linguistic conventions according to which background linguistic conventions may be rapidly expanded to incorporate new word meanings or shifted to revise the meanings of words already in circulation. I argue that this dynamic account of conventions both resolves the problem raised by cases of lexical innovation and that it does so in a way that is preferable to those who—like Davidson—deny important explanatory roles for linguistic conventions. (shrink)

Situation semantics was developed as an alternative to possible worlds semantics. In situation semantics, linguistic expressions are evaluated with respect to partial, rather than complete, worlds. There is no consensus about what situations are, just as there is no consensus about what possible worlds or events are. According to some, situations are structured entities consisting of relations and individuals standing in those relations. According to others, situations are particulars. In spite of unresolved foundational issues, the partiality provided by situation semantics (...) has led to some genuinely new approaches to a variety of phenomena in natural language semantics. In the way of illustration, this article includes relatively detailed overviews of a few selected areas where situation semantics has been successful: implicit quantifier domain restrictions, donkey pronouns, and exhaustive interpretations. It moreover addresses the question of how Davidsonian event semantics can be embedded in a semantics based on situations. Other areas where a situation semantics perspective has led to progress include attitude ascriptions, questions, tense, aspect, nominalizations, implicit arguments, point of view, counterfactual conditionals, and discourse relations. (shrink)

Explicatures are not cancellable. Theoretical considerations.

Paulina Sliwa (2015) argues that knowing why p is necessary and sufficient for understanding why p. She tries to rebut recent attacks against the necessity and sufficiency claims, and explains the gradability of understanding why in terms of knowledge. I argue that her attempts do not succeed, but I indicate more promising ways to defend reductionism about understanding why throughout the discussion.

Algorithms for NLG NLG is typically broken down into stages of discourse planning (to select information and organize it into coherent paragraphs), sentence planning (to choose words and structures to fit information into sentence-sized units), and realization (to determine surface form of output, including word order, morphology and final formatting or intonation). The SPUD system combines the generation steps of sentence planning and surface realization by using a lexicalized grammar to construct the syntax and semantics of a sentence simultaneously.