Linguistics Research Seminar Series
Three factors for the early bias in scope assignment
Dr. Vincenzo Moscati, Università degli studi di Siena
Well before the birth of modern Cognitive Science, the study of ambiguous stimuli has provided many insights about the way our mind organizes the incoming material and the principles that rule its automatic operations. If we consider more recent findings in psycholinguistics, the study of local and global ambiguities have led to the discovery of consistent effects (e.g. garden paths) that are potentially telling about the biases active on the human parser. The same logic can be also fruitfully applied to the study of Language Acquisition: through the study of ambiguous sentences, not only can we observe early processing limitations in young children, but also the existence of learnability issues that might explain the different behaviour between children and adults.
A well-studied ambiguity is the one generated by the presence of two logical operators within the same sentence. Early studies (e.g,, Musolino 2000) focused on the scope interactions between negation and nominal quantifiers (e.g. the detective didn’t find someone/everyone), explaining the difference in the interpretation of these sentences in terms of surface structure (Musolino & Lidz 2006). This account has been challenged in a series of successive studies, showing that other factors, different from surface structure, can account for: i. the absence of inverse scope readings in the early Musolino’s study ii. deviant scope readings with other types of scope-bearing elements (e.g., modals), iii. the overgeneration of inverse scope readings not allowed by the adult grammar. Three factors have been isolated thus far: A) the logical strength of the competing readings, captured in terms of logical entailment B) discourse adherence, modelled in terms of the Question-Answer Requirement C) the type of operators involved, described in terms of the domain of quantification: over possible worlds or over individuals
In this talk I’ll focus on the data coming from modal sentences, showing that until the age of 5, adult readings cannot be consistently accessed despite their adherence to the QAR. With these sentences, factors A) logical strength and C) type of operator overrule the effect of discourse-pragmatics attested with nominal quantifiers. I will also present an argument, based on Moscati & Crain (2014), discussing why strong readings should be preferred over weak ones by young children.
Friday the 28th of November, 2pm – 3pm, Room 17C22
Jordanstown Sports Village Location
Jordanstown Sports Village is the home of sport at Ulster University, located just seven miles north of Belfast.