Search

This paper reports on the state-of-the-art in application of multidimensional scaling (MDS) techniques to create semantic maps in linguistic research. MDS refers to a statistical technique that represents objects (lexical items, linguistic contexts, languages, etc.) as points in a space so that close similarity between the objects corresponds to close distances between the corresponding points in the representation. We focus on the use of MDS in combination with parallel corpus data as used in research on cross-linguistic variation. We first introduce the mathematical foundations of MDS and then give an exhaustive overview of past research that employs MDS techniques in combination with parallel corpus data. We propose a set of terminology to succinctly describe the key parameters of a particular MDS application. We then show that this computational methodology is theory-neutral, i.e. it can be employed to answer research questions in a variety of linguistic theoretical frameworks. Finally, we show how this leads to two lines of future developments for MDS research in linguistics., 40 pages; pre-print; accepted for publication in Corpus Linguistics & Linguistic Theory (Nov. 2021)

This paper surveys the strategies that the Contrastive, Typological, and Translation Mining parallel corpus traditions rely on to deal with the issue of target language representativeness of translations. On the basis of a comparison of the corpus architectures and research designs of the three traditions, we argue that they have each developed their own representativeness strategies: (i) monolingual control corpora (Contrastive tradition), (ii) limits on the scope of research questions (Typological tradition), and (iii) parallel control corpora (Translation Mining tradition). We introduce normalized pointwise mutual information (NPMI) as a bi-directional measure of cross-linguistic association, allowing for an easy comparison of the outcomes of different traditions and the impact of the monolingual and parallel control corpus representativeness strategies. We further argue that corpus size has a major impact on the reliability of the monolingual control corpus strategy and that a sequential parallel control corpus strategy is preferable for smaller corpora.

The variation in distribution and meaning of the English Present Perfect compared to its counterparts in other European languages raises a puzzle for the cross-linguistic semantics and pragmatics of tense and aspect. We apply Translation Mining, a form-based approach, to analyze the meaning of the HAVE-PERFECT across languages in a parallel corpus based on Harry Potter and the Philosopher’s Stone and its translations in Swedish, Spanish, Dutch, German and French. We use the alternation in the HP novel between narrative discourse (storytelling) and dialogue (the characters talking to each other) to establish the PERFECT as an indexical tense-aspect category that appears exclusively in dialogue. We then link the proposed information management roles of the Present Perfect (Portner 2003; Nishiyama & Koenig 2010) to moves in the language game. We find different distributions of PERFECT use across the sentence types corresponding to these moves (declarative vs. interrogative). This lends support to a cross-linguistically common rhetorical structure in sequences of PERFECT sentences (de Swart 2007).

This paper studies a construction in which specificational 'namely' takes a modal expression as antecedent, and an if- or when-clause as argument (example: Working as a filmmaker can be taxing, namely if you're required to get sleek product shots). Such cases do not satisfy previously claimed generalizations about the behavior of 'namely' in Anderbois & Jacobson (2018) and Onea (2016). Moreover, these cases show that modal expressions can raise an implicit question that gets answered by an if/when-clause. Not all types of modals allow this -- the paper argues that it is restricted to Portner's (2009) category of "quantificational modals". Following Portner's (2009) idea that quantificational modals are quantifiers over situations, the central proposal is that a modal or adverbial operator is licensed as antecedent of 'namely' only when the formal objects that are existentially quantified over can be singled out and identified. This leads to a difference in inquisitiveness between situation-quantifiers and world-quantifiers. This account gives insights into the inquisitive character of modal operators, and elucidates the difference between modal restriction by if-clauses (in regular conditionals) and modal specification by if-clauses (in the construction studied here). Finally, it provides a better view of the role of if- and when-clauses in a question-answer framework of discourse.

This dissertation studies counterfactual conditionals in English. Counterfactual conditionals typically come with inferences that their antecedent and their consequent are not true in the actual world (written CF-p and CF-q, respectively).(1) If John had taken the bus, he would have been on time. CF-p: John did not take the bus CF-q: John was not on timeFocusing on CF-q, the dissertation answers the question of what contextual conditions must be met for CF-q to arise, by looking at cases in which the inference gets cancelled. I show that CF-q can be cancelled in a number of different contexts:- in some but not all conditionals that have the words 'also' or 'still' in their consequent;- in a set of conditionals when uttered with a specific intonation contour.Empirically, I show that when 'also' focus-associates with material in the antecedent, it cancels CF-q, but when it associates with material in the consequent it does not. This difference in focus association is reflected in the prosodic properties of the particle 'also' in the conditional.The first step of the analysis is to find a property that unites the different CF-q-cancellation contexts. I argue that what they have in common is their making salient more than one cause for the same consequent (a multiple cause context). In the case of 'also', association with focus determines where focus alternatives are generated, which in turn determines whether a single or multiple causes are made salient.The second step of the analysis is to show that conditional perfection (strengthening conditionals to biconditionals) is a necessary ingredient for the CF-q inference to arise, an idea sketched by Karttunen (1971). The crucial prediction is that in contexts in which no conditional perfection occurs, CF-q is not generated either.Finally, I show that multiple cause contexts do not trigger conditional perfection, which in view of the above-mentioned prediction completes the analysis. Moreover, because conditional perfection has been analyzed in terms of discourse structure and exhaustive answers, we obtain a new set of theoretical tools to study counterfactuality as a discourse phenomenon.

This paper proposes a dynamic approach to the problem of split antecedents, in which a pronominal element refers back to a disjunctive clause. First, I discuss the ambiguities disjunction gives rise to in intensional contexts, in particular their consequences for anaphoric possibilities. I pay special attention to the interaction of future tense and disjunction, showing the role of temporal structure in the availability of anaphora to a split antecedent. Second, I propose a dynamic binding account for split antecedents introducing the concept of `variable complexes'. These are syntactic constructs consisting of two variables that can be bound by two corresponding quantifiers. I implement variable complexes formally as an extension of Dynamic Predicate Logic (DPL; Groenendijk and Stokhof 1991). This provides better empirical coverage and also simplifies the status of disjunction in DPL by giving us an enriched view on the dynamic behavior of connectives.