ABSTRACT
Children have been reported to prefer the surface scope or “isomorphic” reading of scopally ambiguous sentences (Musolino 1998, among others). Existing accounts in the literature differ with respect to the proposed source of this isomorphism effect. Some accounts are based on learnability considerations (e.g., Moscati & Crain 2014), while others invoke pragmatic and/or processing factors (e.g., Gualmini et al. 2008; Musolino & Lidz 2006). The present study investigates whether the isomorphism effect is specific to development or rather is observable in other populations with language processing limitations. We investigated the interpretation of ambiguous sentences containing “every” and negation in 4–6-year-old children, individuals with Broca’s aphasia, and neurotypical adult controls. We observed parallel performance in the children and the aphasic group, with both groups accessing more surface scope readings than inverse scope readings. This finding suggests that the preference for isomorphism may not be specific to acquisition and supports accounts that are not specifically based on learnability considerations—for example, processing accounts along the lines of Musolino & Lidz (2006).
Notes
1 A variety of theoretical approaches have been proposed to capture the distinct readings. One standard approach involves positing a covert movement operation that displaces elements at Logical Form (LF) (see, e.g., May 1977; Fox Citation2000). For present purposes, we remain neutral with respect to these theoretical details.
2 These criteria consisted of: (i) assessment on the Western Aphasia Battery (WAB, Kertesz Citation1982); (ii) agrammatic speech production; and (iii) impaired comprehension of “syntactically complex” sentences on the Verb and Sentence Test (VAST, Bastiaanse et al. Citation2003).
3 Note that the SS reading entails the IS reading; it was therefore not possible to present the items in a context that made the SS reading true but the IS reading false.
4 An anonymous reviewer points out that while the IS reading of the test sentences provides a congruent answer to the QUD, the SS reading provides a stronger assertion that asymmetrically entails that congruent answer. Our goal was simply to ensure that both readings of the ambiguous test sentences constituted good answers to the explicit QUD (they both entailed a negative answer to the QUD). We leave to future study the role of strength and “relative” congruence of possible answers.
5 The QAR model (Gualmini et al. Citation2008) could in principle also be extended to account for the performance of the BA group. We observed, however, that even in the context of an explicit QUD (for which the target sentences were good answers), neither the children nor the BA group performed at the level of the typical adults. We leave to future research a more systematic manipulation of the QUD factor.