Review of Image, Language, Brain
Date: Fri, 13 Sep 2002 22:24:18 -0500 (EST)
From: Robert Port
Subject: Psycholinguistics/Neurolinguistics: Marantz, Miyashita and O'Neil (eds.)
Marantz, Alec, Yasushi Miyashita and Wayne O'Neil (eds.) "Image,
Language, Brain: Papers from the First Mind Articulation Project
Symposium." (MIT Press; Cambridge Mass, 2000). 272 pages. With
introduction by the three editors, 12 chapters. Part I. Language and
Brain. Part II. Image and Brain.
Robert Port, Departments of Linguistics, Cognitive Science and
Computer Science, Indiana University.
This book is based on a conference held in 1998 in Tokyo. The purpose
of the meeting was to "examine recent attempts to unify linguistic
theory and brain science that have grown out of the increasing
awareness that a proper understanding of language in the brain must
reflect the steady advances in linguistic theory of the past forty
years" and to "integrate linguistics and brain science in a way that
has been attempted in the study of the visual system."
This collection of research reviews provides an excellent introduction
to the area, although some previous background in neuroscience would
be essential to evaluate and integrate the arguments. The volume
contains good tutorial material on some behavioral research techniques
and recent neuroimaging methods. Linguists with an interest in
neuroscience will find many provocative ideas in these chapters. At
this point, quite a bit is known about what regions of the brain are
active for various components of linguistic processing - both for
production and perception of speech. Although there is certainly no
concensus view emerging from these papers, there is certainly a great
deal of concrete and stimulating evidence about cognitive skills and
Chapter 1. N. Chomsky `Linguistics and the Brain'. This seems to be
the first serious defense of his early views on the "language
acquisition device" in several decades. Suffice it to say, he gives
up nothing. He begins by taunting neuroscientists for the overly
optimistic views of some on how close we are to understanding brain
functions like language and he scoffs at talk of "emergence" as though
it were some deep new insight when, in his view, the term is just a
mask for our continuing ignorance (and, characteristically, that it is
an idea that was anticipated in scientific discussions during the 18th
Chomsky's basic argument is that cognitive science is in the same
position with respect to biology and physiology as chemistry was to
physics until almost the mid-20th century. In the 18th century, he
says, chemists abandoned all attempts to reduce chemistry to physical
principles and simply began cataloguing phenomena and organizing them
as best they could -- without claiming to even be studying anything
real. Unification of the theory of chemistry to physics didn't come
until very recently after physics was itself radically restructured
(by quantum theory) in a way that allowed chemistry to be unified with
it. By analogy, linguists should just plow ahead without worrying
whether linguistic theories can find physiological interpretation. In
due time, it is brain science that will figure out how to accept and
incorporate the results of modern linguistics (he hopes).
Chapter 2. "Cognitive neuroscience of speech processing," by
David Poeppel and Alec Marantz. This chapter offers an excellent tutorial
on the logic and relative strengths of the four new noninvasive
techniques discussed in many later chapters: functional magnetic
resonance imaging (fMRI), magneto-encephalography (MEG),
electroencephalography (EEG) and positron emission tomography (PET).
Chapter 3. "How infants acquire language: Some preliminary
observations," by Jacques Mehler, Anne Christophe and Frank Ramus.
The authors review many Chomsky-inspired ideas about language
acquisition, eg, regarding "language modules", "the language organ",
the acquisition of "parameters", etc and find his ideas to be
generally supportable at least when compared to the earlier
learning-based views. Still, they acknowledge that there is clear
evidence that statistical properties of the presented corpus are also
employed by language-learning infants.
Chapter 4. "The speaking mind/brain: Where do spoken words come from?"
by W. Levelt and P. Indefrey. Noting that "nothing is more useful in
cognitive neuroscience than a well-founded processing theory" the
authors attempt to break down various language tasks (eg, picture
naming, word reading, word repetition, etc.) into hypothesized
component processes. Then they do meta-analysis on the published
literature that use the various brain imaging techniques to infer what
regions of cortex are uniquely involved in each component process.
The outcome is a set of pictures of cerebral cortex showing which
(rather large) regions should be implicated in which processes. A
later article by Frackowiak, however, offers a word of caution
regarding this method of drawing inferences from such addition and
subtraction of cognitive functions: he notes that that these tasks
have many kinds of interactions. Just because, say, one asks the
subject only to pronounce a word doesn't mean that they do not also
process the meaning.
Chapter 5. "The dependency locality theory: A distance-based theory of
linguistic complexity," by E. Gibson. The author reviews earlier
attempts to explain why some syntactic constructions (eg,
center-embedding as in "The reporter who the senator who John met
attacked disliked the editor") are much more difficult than others (eg,
"The senator that John met disliked the reporter who attacked the
editor" and even "The reporter who the senator I met attacked disliked
the editor"). His scoring mechanism for degree of difficulty of
sentences based on his dependency-locality model is supported by
experimentally obtained measures of "comprehension time" for a set of
Chapter 6. "The neuronal dynamics of auditory language comprehension"
by A. Friederici. This work makes comparisons of brain response for
normal German sentences, "syntactic prose" (analogous to "All wimsy
were the borogroves" that are easy to parse but have mostly nonsense
vocabulary), word lists (using the vocabulary of the normal sentences)
and pseudo word lists (like "were, borogroves, wimsy, all," etc). Both
fMRI (for location) and event-related potentials (for time functions)
are obtained to allow interpretation of brain processing.
Chapter 7. "Neural control of cognition and language" by Masao Ito.
This eminent neuroscientist gives us his bigpicture view of cognition
in terms of control theory (using notions like input, controller,
control object, feedback, etc). He proposes that linguistic behavior
and abstract "verbal thought" can be insightfully understood in these
terms especially by integrating recent discoveries about the essential
role of the cerebellum in language and general cognition. The most
striking novel claim is that " mental activities are analogous to
movements" (p. 158). The cerebellum, he says, provides internal
models for control, both for movement and for verbal thinking, neither
of which depends on sensory feedback.
Chapter 8. "Imaging neuroscience: System-level studies of the
physiology and anatomy of human cognition", by R Frackowiak. This
readable paper reviews results on visual and motor imaging and the
corresponding neural maps obtained in recent years using the four main
noninvasive techniques. Then he comments on how these ideas might
apply to cognitive processes and language.
Chapter 9. "Central control of voluntary movement as studied by
multidisciplinary noninvasive approaches," by Hiroshi Shibasaki. A
review of work investigating the commands (or "will") to execute gross
bodily motions and gaits and to stop these motions. The question is to
find just where this command occurs and when it occurs relative to
movement onset. Different noninvasive techniques are appropriate for
different aspects of the question.
Chapter 10. "Neuromagnetic inverse modeling: Application of
eigenstructure-based approaches to extracting cortical activities from
MEG data," by Kensuke Sekihara, David Poeppel, Alec Marantz and
Yasushi Miyashita. Given measurements from various places on the
skull, how can the precise 3D location of an event within the skull be
determined? This paper reviews mathematical aspects of this issue and
presents algorithms to do the job.
Chapter 11. "Competitive mechanisms subserve selective visual
attention" by J. Reynolds and Robert Desimone. This paper reviews studies
by Desimone and colleagues investigating the neural mechanisms of
visual attention when monkeys see one or several visual targets at the
same time. After presenting data supporting their "biased competition"
model, they present a simple neural network and its dynamical equations
that can account for the experimental results.
Chapter 12. "Origin of visual imagery: Its neurophysiological basis,"
by Yasushi Miyashita. This presents studies of the neural basis of
the "depictive internal representations" (that is, roughly, visual
images) used for solving memory tasks by macaques using methods
employing single-unit recordings (that is, with electrodes inserted in
single cells) from the inferotemporal cortex and other techniques.
Every paper in this book is good -- not a dud anywhere. The editors
endeavored to provide enough background material (and probably enough
editorial intervention with the chapter authors) that the book is
relatively accessible to a broad scientific audience. It may still be
a challenge to the typical linguist, but I recommend it for those
interested in the question of how to approach the neural basis of
language. Although Chomsky's chapter seems to suggest that linguists
need not concern themselves much with the matters presented in the
rest of the book, my view is that Chomsky is wrong. What we need is
not major revisions in biology to allow linguistics to be absorbed as
is, but rather major revisions in linguistics to make it fit with
| ABOUT THE REVIEWER:
About the reviewer Robert Port is a phonetician oriented toward cognitive science. He has a PhD in linguistics from University of Connecticut and a Masters in Linguistics from Columbia. His main research area is in the temporal aspects of speech. In recent research, he has been concerned with the tendency toward periodic patterning of repeated phrases and the evidence this offers supporting a role for dynamical systems in speech production and perception. He edited "Mind as Motion: Explorations in the Dynamics of Cognition" (MITP, 1995) with Tim van Gelder.