We are pleased to welcome Nina Haslinger (MIT, ZAS) for our first meeting of 2025 on Monday, January 13 at 3-4pm in Room 117 of the McGill linguistics department. Nina will be delivering a talk with the title, “Undergeneration problems for structural alternative theories,” which is joint work with Viola Schmitt.
After Nina’s presentation, there will be a reception at Thomson House at 4:15-5:30pm. There is funding to cover snacks and the first round of drinks. We will head over to Thomson House as a group. Join us for post-presentation discussion and to kick off the new semester!
For Nina’s 3-4pm talk on Monday, online participants can join with this link: https://mcgill.zoom.us/meeting/register/16ZiB1ATRMGSUxbkGOJGvw. (Note that it is a different link from last semester’s.)
Here is Nina’s abstract:
Over the last decade, most work on grammaticalized strengthening has followed Katzir (2007) and Fox & Katzir (2011) in adopting a structural constraint on the alternatives that form the input to the strengthening operation: An alternative must be derivable from the prejacent via a series of operations that do not increase structural complexity. For instance, *some* can be strengthened to *some but not all* by negating *all*, but a strengthening to *all* based on the more complex alternative *not all* is not possible. While successful in some empirical domains, this proposal faces well-known problems of both over- and undergeneration (Romoli 2013, Trinh & Haida 2015, Breheny et al. 2018, Schwarz & Wagner 2023, 2024 a.o.). Recent work has therefore proposed to eliminate structural constraints on alternatives altogether (e.g. Hirsch 2024). In particular, Schwarz & Wagner (2023, 2024) propose a condition that blocks a given strengthened meaning for a sentence if there is a sentence of equal or lower structural complexity that expresses the same meaning without strengthening.
This talk takes an exploratory look at the question of which features of the structural complexity approach should be retained despite the undergeneration problems. We start by sketching a more conservative approach to one of the problems motivating Schwarz & Wagner’s blocking theory, namely that alternatives of arbitrary complexity become available in the scope of universal modals. We propose that, instead of comparing the structure of each alternative to the prejacent, two alternatives should be compared to one another in case there is a “stalemate” between them, i.e. in case it is not possible to consistently negate both of them. Since universal modals remove stalemates, they permit the alternatives to be formally unconstrained.
In the second half of the talk, we explore another type of undergeneration problem, first noted by Trinh & Haida (2015): In some cases, stalemates between alternatives can be broken by manipulating what is relevant in the context, while structural complexity is disregarded. While we will not propose a solution, we will make some new observations about the preconditions under which such “context-relative stalemates” occur and use these observations to identify some aspects of the Fox & Katzir approach that should be retained even if the core idea of measuring the complexity of each alternative against the prejacent is given up.