MEG evidence for conceptual combination but not numeral quantification in the left anterior temporal lobe during language production.

Paul Del Prato,Liina Pylkkã¤Nen

doi:10.3389/fpsyg.2014.00524

Abstract

The left anterior temporal lobe (LATL) has risen as a leading candidate for a brain locus of composition in language; yet the computational details of its function are unknown. Although most literature discusses it as a combinatory region in very general terms, it has also been proposed to reflect the more specific function of conceptual combination, which in the classic use of this term mainly pertains to the combination of open class words with obvious conceptual contributions. We aimed to distinguish between these two possibilities by contrasting plural nouns in contexts where they were either preceded by a color modifier (“red cups”), eliciting conceptual combination, or by a number word (“two cups”), eliciting numeral quantification but no conceptual combination. This contrast was chosen because within a production task, it allows the manipulation of composition type while keeping the physical stimulus constant: a display of two red cups can be named as “two cups” or “red cups” depending on the task instruction. These utterances were compared to productions of two-word number and color lists, intended as non-combinatory control conditions. Magnetoencephalography activity was recorded during the planning for production, prior to motion artifacts. As expected on the basis of comprehension studies, color modification elicited increased LATL activity as compared to color lists, demonstrating that this basic combinatory effect is strongly crossmodal. However, numeral quantification did not elicit a parallel effect, suggesting that the function of the LATL is (i) semantic and not syntactic (given that both color modification and numeral quantification involve syntactic composition) and (ii) corresponds more closely to the classical psychological notion of conceptual combination as opposed to a more general semantic combinatory function.

Highlights

A fundamental question for the science of language is how general processes such as lexical access and composition decompose into specific computational subroutines
As regards the combinatory operations of language, brain research on sentence processing has in recent years progressed from broad characterizations of networks of regions implicated for the composition of sentences (e.g., Mazoyer et al, 1993; Just et al, 1996; Xu et al, 2005; Brennan and Pylkkänen, 2012) to the finding that among these regions, the left anterior temporal lobe (LATL) is consistently involved even in the most basic combinatory operations, such as building a small twoword phrase (Bemis and Pylkkänen, 2011a, 2013a,b)
We asked whether the LATL is computationally specialized for the combination of conceptually rich content words, which have constituted the stimulus material of prior basic composition studies (Bemis and Pylkkänen, 2011a, 2013a,b; cf., Baron and Osherson, 2011), or whether its function extends to cases of numeral quantification, which would suggest a more general combinatory role

Summary

Introduction

A fundamental question for the science of language is how general processes such as lexical access and composition decompose into specific computational subroutines. As regards the combinatory operations of language, brain research on sentence processing has in recent years progressed from broad characterizations of networks of regions implicated for the composition of sentences (e.g., Mazoyer et al, 1993; Just et al, 1996; Xu et al, 2005; Brennan and Pylkkänen, 2012) to the finding that among these regions, the left anterior temporal lobe (LATL) is consistently involved even in the most basic combinatory operations, such as building a small twoword phrase (Bemis and Pylkkänen, 2011a, 2013a,b) This allows us to ask a computationally more specific question: what aspect of basic composition is the LATL responsible for?. Behavioral evidence has shown that the conceptual and grammatical encoding of short two-word utterances such as those produced in our study is completed www.frontiersin.org

Objectives

Methods

Results

Discussion

Conclusion