Investigating syntactic and semantic variation in exceptive-additive constructions from a crosslinguistic perspective

• Applicant: Professor Dr. Clemens Steiner-Mayr
• Subject Area: General and Comparative Linguistics, Experimental Linguistics, Typology, Non-European Languages
• Term: since 2021
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 457168471

This project explores the syntax and semantics of exceptive-additive (EA) constructions crosslinguistically. The variants of (1) are instances of such constructions.(1) Every boy but / besides / except Joe came.EA constructions fall into three categories: (i) unambiguously exceptive, (ii) ambiguously exceptive-additive, and (iii) unambiguously additive. On (i) a substantive amount of literature exists. The variants of (1) with “but” and “except” are examples, showing the exceptive meaning ‘Boy Joe did not come and every other boy came’. (ii) and (iii) have received little attention so far despite (ii) being more common crosslinguistically than (i) (and (iii)). For (ii), (1) with “besides” is an example receiving again the exceptive meaning. Crucially, “besides” in (2) yields the additive meaning ‘Boy Joe came and some other boy came’, whereas “but” and “except” are unacceptable.(2) Some boy *but / besides / *except Joe came.For (iii) Spanish “además de” is an example. It is the reverse of “but” and “except” in that it is incompatible with “every” but fine with “some”, where it gets the additive meaning.The project consists of two work packages (WPs) with the following goals. WP1 deals with the considerable variation found both in the (iv) syntax and (v) semantics of EA constructions. This variation has not been systematically studied and is therefore ill understood. WP1 is a large-scale data collection driven by theoretical considerations. It studies both crosslinguistic variation and such within languages. Since WP1 investigates understudied languages it will reveal the true breadth of variation. Regarding (iv), WP1 asks which EAs can precede clausal material and which cannot. E.g., the material in the EA following “besides” and “except” in (1) can be clausal, but not so with “but”. Regarding (v), WP1 asks which exact EA meanings are available in a language and what they depend on. E.g. in (1) and (2) the exceptive reading aligns with “every”, and the additive one with “some”. But in Tundra Nenets both contexts only allow for the exceptive reading.WP2 builds on the data gathered in WP1 to develop a unified formal semantic account of EA constructions of varying syntactic sizes. Since most existing theories of EA constructions build on observations about ”but”, which cannot precede full clauses, they make the wrong predictions for many data. The account built on the basis of the results from WP1 will therefore be descriptively more adequate. It will both capture similarities between languages and model as much of the semantic variation found as possible. The two key ideas underlying WP2 are: the availability of a certain reading of an EA depends on its own logical properties and those of the other functional elements in the sentence; the exceptive-additive ambiguity is structural, i.e., ambiguous EA expressions decompose into two operators with variable scopes, whereas unambiguous EA expressions do not do so.

• DFG Programme: Research Grants