Uniquely Human Language Processing
Kuniyoshi L. Sakai, The University of Tokyo

1. Introduction
Language is a system of association between an infinite set of meanings and sounds (or signs). Modern linguistics has discovered that this system of association is crucially mediated by highly abstract structures. Thus, the human language faculty must comprise, as its core component, a finite system that generates an infinite set of these structures associated with linguistic expressions, each of which is a meaning-sound pair. This generative system is called syntax, and its nature and functioning have been at the center of much linguistic theorizing in the past fifty years or so. The challenging task for systems neuroscience is to determine the physical basis for this system, thereby providing suggestive (hopefully decisive) answers to such important questions as: (i) to what extent syntax is autonomous, i.e., independent from sound (phonology), meaning (semantics), memory restrictions, and other cognitive factors; and (ii) what “syntax” essentially refers to in human language. Some aspects of syntax have been thought to be a uniquely human component of the faculty of language [Hauser et al., 2002; Sakai, 2005]. Fitch and Hauser (2004) have reported that monkeys can master ABAB string patterns but not AABB string patterns, where A and B stand for distinct sets of syllables. In apparent contradistinction with this work, a recent study has reported successful mastery of both string patterns by songbirds after ten thousand trials of training [Gentner et al.,
2006]. These important studies have implicitly made the assumption that the strings are structured as either [AB][AB] corresponding to a finite-state (regular) grammar {S → ABS, S → AB} or A[AB]B corresponding to a phrase structure (i.e., more powerful) grammar {S → ASB, S → AB}, where S is a non-terminal symbol, and
A and B are terminal symbols that will be replaced by actual stimuli. However, quite apart from the finitary status of the data examined in these studies, which could be problematic when viewed from the results of formal language theory, the string AABB could also be interpreted as [AA][BB] with a counting strategy for the repeated stimuli AA and BB. The present study using artificial language experiment for humans is an attempt to clearly dissociate the meaningless string patterns and the hierarchical structures associated with the strings. This research would clarify what aspects should be focused on when human language and animal pattern processing are compared.

2. Results
We performed an artificial language experiment on humans using functional magnetic resonance imaging (fMRI) that involved sentences consisting of written pseudowords and grammatical particles, thereby changing the structure types, string patterns, and memory load of the sentences separately. We found that the activation in two separate regions in the left inferior frontal gyrus (IFG) differentially increased when syntactic analysis was enhanced. More specifically, the results suggest that the left dorsal IFG (dIFG) activation was enhanced by analyzing the syntactic structures of embedded sentences (ES) than simple sentences (SS) and compound sentences (CS), and that the left anterior IFG (aIFG) activation was modulated by applications of the Agree operation. These results from systems neuroscience provide direct evidence for the claim that syntactic structures indeed underlie human linguistic computation, and that the human syntactic ability is independent from such factors as string patterns, general memory load, priming/counting, and semantics.
It is striking to note that the modern linguistic theory based on syntactic structures provides an illuminating explanation for the left IFG activation presented here. The successful combination of experimental findings and theories promises further development of the study of the human language faculty, which could facilitate the true unification of hitherto disparate disciplines in the sciences and humanities.

Fitch WT, Hauser MD (2004): Computational constraints on syntactic processing in a nonhuman primate. Science 303:377-380.

Gentner TQ, Fenn KM, Margoliash D, Nusbaum HC (2006): Recursive syntactic pattern learning by songbirds. Nature 440:1204-1207.

Hauser MD, Chomsky N, Fitch WT (2002): The faculty of language: What is it, who has it, and how did it evolve? Science 298:1569-1579.

Sakai KL (2005): Language acquisition and brain development. Science 310:815-819.

# Uploaded 181 Plays 0 Comments


Kavli Frontiers of Science PRO

This channel contains session presentations that cover biological topics from the Kavli Frontiers of Science symposium series of the National Academy of Sciences.

For additional symposium information, please visit our web site (

Browse This Channel

Shout Box

  • Julio Molina Montenegro
    What shapes the questions asked by scientific researchers?
    As part of the series A Portrait of a Question, this short documentary explores Dr. Kevin McGhee's interest in schizophrenia and how fruit flies can help identify some of the genetic factors behind this brain disorder.

    by Julio Molina Montenegro

Heads up: the shoutbox will be retiring soon. It’s tired of working, and can’t wait to relax. You can still send a message to the channel owner, though!

Channels are a simple, beautiful way to showcase and watch videos. Browse more Channels.