Cortical representation of quantification: The role of the left anterior temporal lobe

Quantifiers like 'all' or 'some' play a key role in how we construct meaning from language, but their underlying neurobiological mechanisms are still unclear. One difficulty in determining the neural correlates of quantification is that the interpretation of these elements depends on multiple elements of a sentence, unlike other determiners, such as 'the.' This makes it difficult to isolate neural signals that are specific to the logical properties of quantifiers but orthogonal to the intricacies of predictive processing. In this magnetoencephalography study, we overcame this challenge by employing rapid parallel visual presentation of four-word quantified sentences as well as sentences containing the semantically simpler definite determiner 'the.' Since the left anterior temporal lobe (LATL) has been hypothesized to perform Boolean set intersection on conceptual representations, we tested whether activation in this region would also reflect the degree of intersection performed by the quantifier (maximal for "all," intermediate for "some," minimal for "no"). Instead, we found that LATL activation corresponds to whether a determiner involves the representation of a unique domain of individuals as well as whether the expression is referential. We argue that this pattern offers neural support for the theory of Restricted Quantification, which states that the arguments of a quantifier are mentally represented in distinct ways. Overall, these results suggest that the LATL performs conceptual combinatory operations that interact with a model of the discourse. SIGNIFICANCE STATEMENT: A remarkable property of language is that we can produce and comprehend an infinite set of complex meanings from a finite set of building blocks. These building blocks vary widely in how they contribute to higher levels of meaning, and we can exploit their differences when studying the neural mechanisms that underlie our recursive ability to build structured meanings. This study uses brain data to study how our minds form basic logical representations out of linguistic elements known as "quantifiers". Our findings suggest that the left anterior temporal lobe is involved in computing the referential domain of a quantified phrase, and they further have bearing on our representational theories for quantifiers.