AUTHOR: Moro, Andrea
TITLE: The Boundaries of Babel
SUBTITLE: The Brain and the Enigma of Impossible Languages
SERIES: Current Studies in Linguistics 46
PUBLISHER: MIT Press
María del Carmen Horno-Chéliz, Department of General and Hispanic Linguistics,
University of Zaragoza, Spain
This book aims to build a bridge between cognitive neuroscience and linguistics.
Its author is a well-known theoretical linguist as well as the coordinator of
the Theoretical Linguistics Unit at the Centre of Cognitive Neuroscience
(University ''Vita-Salute'' San Raffaele), thus, his academic endeavors have
mainly focused on the link between these two disciplines and therefore, and as
Noam Chomsky says in the foreword of the book, he is among the most suitable
scholars to write on it.
This book is organized into three independent self-contained chapters. The first
one offers a complete introduction to theoretical linguistics, more specifically
to generative syntax; the second chapter describes two examples where
neuroimaging technology is used to test - from a biological point of view - some
theoretical linguistic hypotheses, and the third chapter puts forward a
speculative hypothesis on the existence of syntactic rules.
Chapter 1. Hidden Texture
After a most pleasurable prologue, the first chapter of the Boundaries of Babel
is a summary of the main findings of theoretical linguistics in the last twenty
Itt should be borne in mind that the author is a generative linguist. Thus, when
he talks about language and the properties of languages, he understands ''natural
languages'' as the biological result of a human specific capacity. In this sense,
this book is actually a bridge between a way to understand language (and
linguistics) and cognitive neuroscience. Moro does not deal in this book with
anything but the human capacity to acquire and use languages from a generative
point of view.
This chapter consists of three sections. ''Methodological issues'', the first one,
is an enlightening summary about how scientific work is undertaken. It is used
as a justification of his point of view: no scientist can tackle all aspects of
reality, so it is absolutely necessary to reduce one's object of study - and
Moro chooses syntax; more specifically, the limits of syntax variation - and
one's theoretical issues - and he chooses generative principles. He also
mentions the concept of ''error'' in linguistics, in a possible attempt to
distance himself from the normative point of view. For the author, the most
interesting ''errors'' are those which are impossible, those no native speaker
ever makes. These ''impossible errors'' are interesting for him, because they can
be analyzed as the limits of language variation. This section finishes
advocating formalism; his defense of formalism is based not just on its clarity
but also on the way formalism can help to find deductive consequences of any theory.
''A Sample of Syntax'', the second section, is a summary of syntax properties:
linearity, discreteness, recursion, dependence and locality.
This is a very good presentation of generative grammar in the last decades. I
would recommend it, therefore, to anybody who wants to start the study of
natural language properties, namely linguistic students and any researchers whom
this theoretical framework is not known. This introduction to syntax provides a
valuable state of the art, and his didactic abilities make it accessible to
non-experts as well. A slight negative point could be that, from time to time,
the author seems to forget this didactic purpose and outlines some technical
problems, which may be too complex for the general public, and are not always
relevant for the general state of the art. A case in point could be the note 49
on page 70, where the author mentions c- and m- command, without explaining them
in detail and without a clear purpose. Nevertheless, the readers will be able to
easily follow the argumentation.
Finally, ''The Ark of Babel'' is the last section of the first chapter and it is
an introduction to linguistic variety in which Moro emphasizes the idea that
properties of languages and limits in variation are neither logical nor
necessary from a communicative point of view. Hence, syntax is as it is for no
important external reasons. Moro compares linguistic variety and biological
variety. In this sense, if differences between biological species can be
understood as differences in their DNA, differences between languages could be
also understood as differences in a few parameters. This proposal can explain
the way children acquire their native language (so quickly and so well, in spite
of the poverty of linguistic stimuli they are exposed to). If parameters theory
is right, children do not have to learn every linguistic rule, they just have to
parameterize their linguistic universals. The author labels this proposal, the
''tabula praeparata'', to distinguish it from both the empirical one (''the tabula
rasa'') or the pure rationalist one (''the tabula inscrita''). This proposal would
also explain the differences between ''pidgins'' (artificial communication
systems, born from two different languages) and ''creoles languages'' (real
natural languages, inside natural linguistic limits, the result of children's
Chapter 2. Language in the Brain
In this chapter the hypothesis of the neurological basis of the limits in
language variation is tested. In particular, Moro's purpose is to discover
whether there is a neurological correlate to the absence of syntax rules based
on linearity. The author believes that it will be the case. Linguistic proofs
(such as comparing languages or describing acquisition processes) are sufficient
to accept this biological basis. Nevertheless, an external (biological) proof
could also be a solid argumentation in favor of it.
To that end, Moro offers an introduction to neurosciences from its origins. He
underlines the discovery of neuroimaging techniques, which can be used with
healthy individuals, as they are not harmful to human beings. He describes
Wernicke's model basis (he starts with Broca's research), but he interprets it
in its own limits, that is, he does not consider any direct link between large
areas of brain and complex human activities; Moro is certain that the
relationship is much more complex than used to be thought one century ago.
This chapter is organized into four sections. In the first one, entitled ''Seeing
Thought'', Moro explains the anatomy of the brain in simple terms. The author
reviews how the brain has been seen by scientists through history until the
origin of electrical behavior of neurons was proved to be in the cortex.
Here, the reader may miss a mention to Camillo Golgi and Santiago Ramon y Cajal.
They shared Novel Physiology or Medicine Prize in 1906, in recognition of their
work on the structure of the nervous system. This is even more striking if we
realize Andrea Moro is Italian, as Golgi was.
Moro describes two neuroimaging techniques: Positron Emission Tomography (PET)
and functional Magnetic Resonance Imaging (fMRI).
This technical description is less didactic than other parts of the book and the
result can thus be a bit confusing. He does not mention some aspects which are
usually included in non-expert introductions to this technique - for example, in
the description of PET, the use of substances like glucose (e.g.
2-dexosiglucose) or its poor spatial resolution. In addition, he presents
technical details not usually described in popular literature (e.g. Pinel, 2006,
or Vendrell, Junqué and Pujol, 1995). In any case, leaving aside the technical
description of these instruments, Moro clearly shows their most interesting
characteristic: the fact that they can describe cerebral activity, or, more
specifically, they can signal the zones in the brain activated while any task is
The research method he uses, ''Subtractive Method'', is really appropriate and its
appropriateness lies in its simplicity. It consists of visualizing individuals
doing different tasks, so that each test is different from the last one just in
one property. The idea is that when subtracting common areas, the result is the
specific area implied in the specific property. In this section, Moro gives us
again a lecture of what a scientific work consists of. He is aware of his own
limits and he explicitly states that it is not possible to control every
neuronal activity in experimental individuals - it is not possible to be sure
whether the experimental individual is thinking about (or feeling) something else.
In ''The autonomy of Syntax: How to Fool the Brain with Errors'', the second
section in this second chapter, Moro explains the details of his investigation:
the kind of individuals who were chosen (eleven Italian right-handed men) and
the kind of task they had to do. They were given some sentences, which were not
Italian sentences (words were invented, they did not have any meaning) but they
obeyed all Italian linguistic rules (phonological, morphological and syntactic)
but one. Thus, experimental individuals spot the kind of mistake (phonologically
impossible words in Italian, problems with agreement or errors related to linear
word order). Imaging techniques and the subtraction method are used here to
check whether each level activates a different zone in the brain. In this
research it is interesting to take into account that syntax activates one
different zone than other levels do. This result is compatible with the formal
linguistic theory of autonomy of syntax. It is not as relevant to determine
which specific area is activated as to determine that there is actually a
specific syntax zone.
The third section is entitled ''Possible Grammar versus Impossible Grammars''. In
other sections of the book it is claimed that some syntactic rules are universal
(that is, they are present in all human languages). However, here the aim is to
test if there is a neurological basis for these rules. In this case, German
speakers are taught some Italian and Japanese syntactic rules. Some of these
rules are real, but others are not. In fact these last rules are ''impossible
rules'', in the sense that they are not applied in any human language (they are
based on linearity, instead of on hierarchy). As in the previous section, the
imaging techniques and the subtraction method show that different zones are
activated depending on whether linguistic rules or non-linguistic rules are
being used. This result is very insightful, because it is compatible with the
formal linguistic hypothesis of language autonomy. As Moro says, this result is
also important to suggest that adult second language learning also uses the
limits of Universal Grammar (these limits would not be applicable, therefore,
just to child acquisition).
On the whole, it could be claimed that this research is not a real scientific
experiment, in the sense that the causal relationship between language behavior
and neuronal activation has not been proved. The author is, nevertheless, aware
of it and this research is suggestive enough to start a research path that other
experimental tests could take. It is difficult to control all the variables in
this kind of experiment. In my opinion, the only way to control them is to rely
on a statistically relevant number of individuals, so that variables could be
controlled. However, it is obvious that nowadays using neuroimaging technology
is expensive and difficult. Another possibility would be to use Transcranial
Magnetic Stimulation, but there are not enough tests to be sure of its effects
on human health. The future will tell us what the best way to proceed is.
The last section in this chapter is ''Why Isn't There a Mendelian Linguistics
(Yet)?''. The author focuses on the relationship between genes and language. Moro
warns about this following fact: in spite of what many researchers think, the
generative program does not exactly call for a genetic basis for language. Moro
accepts that, if one property is common to every member of one species and it is
alien to any other species (such as language is), this property must be genetic.
However, the existence of a gene of language is not a very plausible idea for
many reasons. There does not seem to be a family (genetic) malformation which
affects just syntax (the so called FOXP2 gene generates not only language, but
also other problems, e.g. motor problems). The author suggests the possibility
that the linguistic genetic basis is pleiotropic, that is, that the interaction
of some genes has multiple consequences and language is one of them.
Chapter 3. The Form of Grammar
In this third and last chapter, the author provides a two-level argumentation
both as regards complexity and purpose: 1) from a general point of view, the aim
of this chapter is to demonstrate that principles that restrict language are
neither logical nor historical nor cultural nor cognitive; these principles are
the result of physical and biological environment of language; it is thus
related with the rest of the book and it is addressed to every possible reader,
regardless of their specific training; and 2) as an illustration of his
proposal, he presents a new theoretical hypothesis; it is thus aimed only at
The chapter is structured in two sections. The first one, entitled ''Logical
versus Learnable, or why Do Languages Have Rules?'', states some reasons for the
existence of syntax (syntax is seen here as one clear limit to linguistic
variety). At this point, Moro is interested in denying the teleological nature
of language. As he points out, language characteristics are not to be more
effective (e.g. in their communicative purposes), but they are as they are due
to their evolution. At most we can say that, if they exist, they must have
reached an equilibrium between their advantages and their disadvantages. They
are just one possibility compatible with the environment. The advantages are
related to language acquisition, whereas the disadvantages are connected with
the fact that syntax limits our communication possibilities.
The second section, entitled ''On the Linear Nature of the Linguistic Signal, or
Why Do Languages Have exactly these Rules?'' focuses on the nature of language:
its properties must be the result of environmental pressures. There are a lot of
possible pressures in environment, but there is one Moro selects to analyze in
depth: linearity of sound. He uses a very up-to date theory to exemplify this
pressure: Antisymmetry Theory by Kayne (1994). This theory has been modified by
Moro in a light version (the Dynamic Antisymmetric Theory).
Antisymmetry Theory predicts that one linguistic element precedes another one
iff it belongs to a phrase which is more prominent than the one the second
element belongs to. This proposal is a very restricted one and it predicts that
the only linear word order could be: Specifier- Head - Complement. The advantage
of this theory is that it simplifies language data: it explains why there are no
languages with inverse word order, that is, Complement-Head-Specifier,
explaining, also, why there are languages which present mixed systems, as the
result of a movement. This is the case of German, where every phrase has the
expected order (Specifier-Head-Complement) but Tense Phrase, which seems to be
another linear word order (Specifier-Complement-Head). In this case, Tense
Phrase in German should then be the result of a movement.
Linguistic Movement is another linguistic principle neither logical nor
necessary. In the standard generative model, it was justified as a way to reach
a full interpretation of phrases (getting rid of non-interpretable features).
This theory generates important questions, such as what exactly are
non-interpretable features? or why do non-interpretable features exist at all?.
Moro suggests an alternative proposal to linguistic movement: For him, movement
is the result of Dynamic Antisymmetry Theory. As a light version of Antisymmetry
Theory, it lets structures be symmetric before the spell-out and it is at this
moment when, sometimes, movement is necessary (in order to break a symmetry
point). In this new proposal, morphology is seen as a filter for which elements
can be moved and where.
This proposal is supported by the existence of intermediate possibilities in
word order or by the ''equative sentences'', where both phrases could be in both
places (before or after the copula), showing a ''point of symmetry''. This is
clearer in Italian, where the second phrase is the one whose characteristics are
those of a subject. Nevertheless, other syntactic contexts seem to be better
explained by the Standard Theory of Movement. Therefore, Moro does not give an
answer to this controversy, but states that a proposal that encompasses both
theories should be made. Perhaps he does not provide such a proposal because it
lies outside the scope of this book. He introduces it just as an example to show
how environmental physical aspects have decisive effects on language properties.
This is a really insightful book both for neuroscientists (or neuroscience
students) interested in language and for linguists (or linguistic students)
interested in neuroscience. It could be a good reference also for the general
public interested in the relationship between the brain and the diversity of
languages. This book is written in such a way that it surely will be appealing
for both researchers and non-experts in the field. This is a unique achievement
accomplished by the author's obvious purpose of being clear in his arguments,
his excellent writing and a clear structure: the three independent
To sum up, this suggestive book is excellent to introduce oneself to the
possible link between theoretical (generative) linguistics and neuroscience
cognitivism. It is organized in such a way that every reader can choose which
chapter(s) to read according to there training and interests. However, its
linear and full reading is a pleasure. As regards the topic of the book, one
conclusion can be drawn: nowadays it is too soon to understand the real link
between the brain and language, but it indeed opens up a fascinating road ahead
Kayne, R. (1994) _The Antisymmetry of Syntax_. Cambridge, MA: MIT Press.
Pinel, J. P. J. (2006) _Biopsycology_. Pearson Education Inc.
Vendrell, P., C. Junqué and J. Pujol (1995) ''La Resonancia Magnética funcional:
una nueva técnica para el estudio de las bases cerebrales de los procesos
cognitivos''. _Psicothema_, 7.1: 51-60.
ABOUT THE REVIEWER:
Dra. Horno-Chéliz is Associate Professor in the General and Hispanic Linguistics
Department at the University of Zaragoza (Spain). She is a member of Sylex Group
Research, focused on the link between syntax and lexical information. She is
also interested in logical semantics. She is currently exploring the link
between language and the brain and its application to second language learning.