Publishing Partner: Cambridge University Press CUP Extra Publisher Login
amazon logo
More Info

New from Oxford University Press!


Style, Mediation, and Change

Edited by Janus Mortensen, Nikolas Coupland, and Jacob Thogersen

Style, Mediation, and Change "Offers a coherent view of style as a unifying concept for the sociolinguistics of talking media."

New from Cambridge University Press!


Intonation and Prosodic Structure

By Caroline Féry

Intonation and Prosodic Structure "provides a state-of-the-art survey of intonation and prosodic structure."

Review of  The Prism of Grammar

Reviewer: Annette E. Hohenberger
Book Title: The Prism of Grammar
Book Author: Thomas Roeper
Publisher: MIT Press
Linguistic Field(s): Syntax
Language Acquisition
Subject Language(s): English
Issue Number: 19.1285

Discuss this Review
Help on Posting
AUTHOR: Roeper, Tom
TITLE: The Prism of Grammar
SUBTITLE: How Child Language Illuminates Humanism
SERIES: A Bradford Book
YEAR: 2007

Annette Hohenberger, Middle East Technical University (METU), Ankara, Turkey

Tom Roeper resolves the methodical challenge of writing an accessible book for
laymen by providing instructions for numerous ''explorations'' which he encourages
the readers to carry out themselves, ideally with their own children. Roeper's
main theme, the humanistic approach to cognitive sciences, goes beyond
linguistics and has its proper place in philosophy, especially ethics. This
global philosophical topic finds its strongest expression in the passionate plea
that parents, pedagogues, and cognitive scientists should preserve the dignity
of children when evaluating their linguistic and cognitive achievements. In an
attempt to bring insights in cognitive science to bear on pedagogical and
political decisions, he invokes ''modularity'' as a suitable theory of the human
mind. Modularity guards against unwarranted generalizations from a local, e.g.,
linguistic, deficit to a general cognitive deficit, thus naturally preserving
the dignity of the child.

This is Tom Roeper'most recent book on language acquisition and on the
background of modern generative language acquisition studies (p XV). He uses the
metaphor of a prism since much in the way light is dispersed by a prism, the
stream of sounds is analyzed by the grammar and different information is
relegated to various modules such as memory, vision, emotions, and intentions (p
3). For the researcher, language, more precisely, grammar, acts as a laser into
the life of the speakers, allowing insights into the human nature (p 3). Roeper
intended to write ''a book that is at once light and serious, that mixes depth
and delight, that is open to abstract and unresolved philosophical issues but
concrete enough to offer parents, teachers, and children immediate experience
with grammar'' (xii). Taking into account that all knowledge entails
responsibility, the book reflects ethical and moral implications of how we talk
to and about children and how we can treat them with dignity (p 4).
The book rests on three pillars: (1) formal principles of language acquisition,
(2) dialectal variation, and (3) social, educational, philosophical implications
of human grammar. It is organized into four major parts: I Goals and Grand
Perspectives; II Why Language Acquisition Is a Challenge to the Child; III
Microdialects and Language Diversity; IV Finding Philosophy and Morality in
Every Sentence

Part II ''Why language acquisition is a challenge to the child'' is partitioned
into six chapters (chapters 3-8) that deal with the acquisition of first words,
first phrases, reference, recursion, ellipsis, and plurals.

In chapter 3 ''First words: Glimpses of the mind'' Roeper presents new
exemplar-based evidence on the understanding of expressive words such as ''uh-oh''
or ''oops'', the use of pronouns and the greeting particle ''hi'', with the
intention of showing what these first words reveal about the early mental life
of children. ''Uh-oh'', for example, is used when something unexpected and
unpleasant happens. Its proper usage presupposes that the child can imagine a
state of affairs that is different from the current one and compare these two -
a rather sophisticated mental act. ''Oops'', on the other hand, presupposes an
agent and refers to an act in the immediate past - hence the concepts of
'agency' and 'time' are involved. As for the proper use of pronouns such as
''it'', children have to infer their meaning and role in discourse from elliptical
utterances of parents. The shared knowledge that is built up in joint
communicative situations casts doubt on the assumption that young children are
egocentric. However, adults and children may differ on their conception of what
''it'' refers to: adults understand ''it'' as an object whereas children seem to
first understand ''it'' as an activity as in ''stop it'' or ''do it'' (p 42).

In chapter 4 ''First phrases: Glimpses of grammar'' Roeper explores the onset of
grammar in child language. The expression of relations is fundamental for the
emergence of syntax as a formal device for linking two entities. Generally,
thematic relations such as 'agent', 'object', 'possession' or 'goal' are
preferably encoded through case, agreement or word order, whereas less abstract
or relevant relations are expressed periphrastically, e.g., through a
prepositional phrase. Interestingly, children do not start out with a seemingly
easy relation such as ''and'', hence we do not hear clauses like ''mommy daddy'' but
they rather start with verb-object relations such as ''pick glove'' or
possessor-object relations such as ''mommy sock'' (p 53). The reason lies in the
basic asymmetry of syntactic structures: one of the two elements is the head
which also determines the category label of the newly formed constituent, the
other its argument or modifier. Such asymmetric syntactic relations between
words are a basic organizing principle of human language. Since coordination
(''and'') is symmetric, it appears rather late. Formally, putting words together
is captured in the notion of 'merge' which comprises linking as well as
labeling. Once 'merge' is mastered, children show ample evidence for recursion
as in multiple compounding, for the playful elicitation of which Roeper gives
instructions in a series of explorations. Parents and children may engage in
delightful recursive compound games where e.g. ''animal cup'' is expanded into
''animal cup house'' or ''print-maker'' into ''print-maker maker'' (p 60). With
recursion the child has the syntactic device for unlimited linguistic
generativity and creativity in her hands.

Chapter 5 ''The absence of absolute reference'' deals with referential expressions
such as ''it, that, there, here, not, no'' and quantifiers like ''all, every, a,
some, the''. These words are particularly suited to show that in human language,
as opposed to animal language, there is no absolute reference. Whether children
appreciate the variable reference of those words above can be tested in a little
exploration where the adult touches her own nose and asks the child: ''Put your
finger here.'' (exploration 5.3, p 69). A child for whom ''here'' has absolute
reference will put her finger to the adult's nose whereas a child for whom
''here'' has variable reference, will put her finger to her own nose. In an
extended series of explorations, Roeper shows how an indefinite ''a'' turns into
''the'', how ''the'' functions as a part-whole indicator or refers to a member of a
set, what the difference between ''not'' and ''no'' and between ''each, every'', and
''all'' is, how negation and quantifiers are combined differently (as in ''not all''
vs. ''all not''), and why young children use multiple negation (as in ''No, I am
not a nothing boy'' (p 93f)). Each of these carefully chosen explorations reveals
some abstract principle of language and tell us something profound about how
language acquisition functions. The absence of absolute reference is fundamental
to human language. Understood as freedom of reference it allows for creative use
of language: We can express and generate new experiences - a freedom inherent in

Chapter 6 ''The heartbeat of grammar'' focuses on recursion, the unique capability
of humans to embed linguistic structures into each other and thereby create
possibly infinite linguistic expressions. In a series of case studies on
increasingly more complex recursive possessor constructions Roeper identifies
various stages in the acquisition process. In the first stage (i) children may
have an understanding of the concept of 'possession' while they may not yet
have a morpho-syntactic analysis available. Only bare possessive pronouns such
as ''mine'' and ''your'' are accessible at this stage. Affixing an '''s'' to a noun
would be evidence for the second stage (ii), as in ''mommy's sock''. In the third
stage (iii) the possissive '''s'' would be affixed to an entire phrase, e.g. the
noun phrase ''[the boy on the corner]'s hat'' (p 121). Embedded possessors such as
''[[Cookie monster]'s sister]'s picture'' will not yet be parsed recursively but
rather understood as a coordination ''a picture of Cookie Monster and his sister''
(p 118). A child having fully mastered recursive possessors will enjoy producing
cascades of embedded possessors such as ''John's friend's dog's hat'' (p 124) and
correctly point to ''the person next to your hat'' as opposed to ''the person next
to you's hat'' (p 112). Finding out where in the grammar recursion is hidden is a
methodological task for the child in language acquisition as it is for the
linguist studying an unknown language. It is in ''these tiny creative engines'' (p
113) - how Roeper refers to recursion in their various linguistic domains -
where the generative power of language lies.

Chapter 7 ''The structure of silence'' is devoted to ellipsis. Ellipsis poses a
challenge to the child since it is structurally as complex as the full utterance
although an element is missing in it. The child who is asked ''I've got peas.
Want some __?'' (p 127) must reconstruct the missing element ''peas''. Ellipsis
shows the automatic, mechanical quality of the grammar which substitutes exactly
the right-sized element that has been elided. The child must figure out what
natural elements in her native language can be elided and which cannot. In a
context where a boy and a girl each have a toy car, when asked ''A girl pushed
her car and so did a boy. Show me'', children around age four reconstruct only
the bare noun (and make the boy push his car) whereas adults and older children
reconstruct the pronoun + noun (and make the boy push the girl's car). The
systematicity of the absence can even help children discover natural units in
their native language. Confronted with the sentence ''John can sing and Bill can
__'' (p 139), they may conclude that English has a natural category AUX since the
modal ''can'' has been stranded and only the full verb ''sing'' has been elided.
Ellipsis shows (i) how UG functions (only natural units can be elided), (ii) how
languages differ (they can elide units of various size) and (iii) how individual
speakers differ (people may elide different micro-features).

Chapter 8 ''The pantheon of plurals'' is a pivotal chapter in that Roeper
discusses the acquisition of plurals and certain problems related to it in an
ethical perspective, as indicated by the subtitle ''from possible worlds to the
ethics of our world''. Plurals, according to Roeper, are the ''DNA of grammar'' (p
160) since they allow us to group tokens of the same type in an act of
abstraction. Children from two years onwards take efforts to make such
abstractions. However, they may differ from those of the adults, as in ''two
hamburger with cheeses'' as opposed to ''two hamburgers with cheese'' (p 166).
Plurality also occurs in double questions where the question word groups
together objects in the world that have to be exhaustively enumerated in the
answer. Hence, in the answer to the question ''Who is eating what?'' (exploration
8.7, p 180) all pairs have to be enumerated exhaustively: (''John is eating
apples; Mary is eating bananas; ...''). However, in some early stages of language
acquisition the exhaustive feature may still be missing. Double questions belong
to a set of core diagnostic tools gathered in the ''Diagnostic evaluation of
language variation'' (DELV). The DELV is a dialectal test including features of
language that are (i) so fundamental that they do not vary across dialects and
(ii) that are indicative of language disorders if they are not mastered by a
child. Roeper explains that the exhaustive reading of wh-words that requires the
notion of a variable is linked to an ''invisible 'every''' housed in the feature
matrix of a wh-word such as ''who'' [wh + someone + every] which enforces the
exhaustive reading. Young children and language-impaired children do not have
this hidden ''every''. Here, the ethical perspective comes into play. Instead of
stigmatization of children Roeper pleads for insightful linguistic intervention
such as presenting the child with wh-questions which - by the semantics of the
verb involved - require two noun phrases in the answer, as in ''who played
together?'' - ''John and Bill'' (p 190). This approach differs from a classical
cognitive approach, sensu Piaget, for whom quantification was a general
cognitive problem. Consequently, if a child did not show proper quantification
she was diagnosed as having a cognitive deficit. Such practice Roeper strictly
refutes, on the grounds of his modular approach to cognition. For him, the
deficit is a grammatical one, more precisely, one of finding out where in the
grammar and by which means cognitive constructs such as 'variables' are
expressed. Claiming a cognitive deficit would require evidence for the same
deficit in other cognitive domains, too. He warns against leaping from language
to cognition, not only for scientific but also for ethical reasons. Diagnosing a
child as cognitively deficient has far-reaching educational and personal
consequences that violate the dignity of the child. Such bad practice can be
avoided by following the advice ''... not to extend a judgment about the
character of a mental module to another module.'' (p 195). Modularity as a
scientific account of the human mind directly translates into ethical behavior
since it does not assume a general cognitive deficit and therefore preserves the
dignity of the child. I will address this point in more detail in the evaluation

Part III ''Microdialects and language diversity'' is a shorter part with three
chapters (9 to 11).
In chapter 9 ''Language variation'' Roeper takes a look at language from the
broader perspective of language variation, especially dialects. As for dialects,
social attitudes quickly dominate our concept of grammar and ''emotion overtakes
structure'', as the subtitle says. Upon listening to a dialect speaker, an
immediate judgment, usually a pejorative one, on the speaker's social class and
identity is made. What starts out neutrally as sensitivity towards language
variation, as in children, becomes a social ruler as soon as they become aware
of social distinctions. The normativity in teaching grammar fosters prejudices
against dialects. In the debate on African American English (AAE, Ebonics)
different points of view on dialects hit upon each other, often irreconcilably.
Roeper counters the layman's view that AAE is a rudimentary and immature
language arguing that AAE is a proper language whose speakers are bilingual. On
a deeper level, we are all bilingual and have multiple grammars stemming from
various language families, a view that Roeper (1999) has dubbed ''Universal

This view of Roeper's is further elaborated in chapter 10 ''Are we all
bilingual?''. Children, having to decide whether a given construction is English,
have to distinguish whether it is related to a general rule or to a particular
lexical item. In any language, aspects of many other languages are lurking
through, hinting at different organizing principles, all made available by UG.
While the core language has rules pertaining to all words of a category, rules
of another language only manifest themselves in few words. The reason why they
survive is that they still rely on rules, albeit from other grammars where they
are productive. A pertinent example of ''multiple grammars in syntax'' (p 214) is
the remnant Verb-Second (V2) rule in English, as in '''Nothing', said John'' (p 215).

Whether languages carry worldviews Roeper discusses by confronting the two big
systems for expressing events, namely tense and aspect. While tense systems
distinguish, e.g., past, present, and future tense, aspect systems distinguish
inceptive, resultative, progressive or static aspect. Whether children are
sensitive to this great divide can be explored in little scenarios where they
have to match pictures with two sentences expressing either resultative aspect,
as in ''Someone is painting a house grey'' vs. ongoing action, as in ''Someone is
painting a grey house'' (p 222). That English combines tense as well as aspect
markers in sentences such as ''John is (tns) runn-ing (asp)'' shows that two kinds
of grammars can be expressed in one language.

Chapter 11 ''The riches of African American English'' takes up the topic of
dialectal variations of chapter 9 and places it in the context of universal
bilingualism of chapter 10. Roeper argues that while AAE looks very different
from mainstream American English (MAE) both languages move along a common
developmental trajectory, only they have arrived at different points on their
journey. Thus, the same rule that allows contraction of the copula ''is'' in MAE,
as in ''John's tall'' has developed further into a deletion rule in AAE, as in
''John tall'' (p 228). In both languages, the syntactic context for the
application of the rule is identical so that contraction/deletion is allowed
sentence-medially but not sentence-finally, hence *''I know how tall
John's/John_'' is illicit in both languages (p 228). Following up on the
principled distinction between tense and aspect systems in chapter 10, Roeper
shows that AAE develops into a rich and recursive aspectual event structure
grammar. The property of AAE that elicits most social negativity is certainly
double negation as in ''No game don't last all night'' (p 235). Again, Roeper
shows that AAE simply sits on a different branch in the universal system of
negation. It applies ''negative concord'', i.e., all negators agree in their [neg]
feature rather than ''cancellation negation'' as does MAE, where each negative
marker cancels out the previous one.
Following the maxim that education should enrich and not control, Roeper advises
against correcting AAE speaking children at school and opts for an education
free from language prejudice which preserves the dignity of childred raised in a
unique language community.

The fourth and last part of the book ''Finding philosophy and morality in every
sentence'' is comprised of four chapters (12-15) in which Roeper addresses
philosophical aspects and ethical consequences of the study of linguistics and
language acquisition, culminating in a plea for the ''defense of dignity''
(chapter 15).

In chapter 12 ''Philosophical consequences: The path from mathematics to human
dignity'' Roeper contemplates whether there exist common principles operating
across the human mind in the sense of a generative grammar. He claims that
''...the 'generative power of grammar' is found throughout the mind/brain'' (p
244). The various cognitive domains share commonalities, e.g., being
hierarchically organized. However, they are spelled out differently in each,
according to innate principles. The working of the mind is instant and
unconscious so that thoughts originating in one module immediately impact on any
other module which has an interface with the former. Roeper summarizes his ideas
in four main points (p 249): ''1. All rapid action or thought must be mechanical.
2. Any mechanism works by principles that can be captured in a mathematical
formula... 3.Each individual has unique formulas. 4. Inner formulas that
generate unique actions cannot be fully comprehended by another person ...''
Derived from those personal formulas is the sense of individuality which calls
for respect of each other's dignity (p249). This is the deepest essence of humanism.

Counteracting rapid, mechanistic, modular thought is ''slow thought'' which
operates on larger time scales such as weeks, months, even years. This ''slow,
deliberate thinking reflects who we are most distinctively... slow thought is
really where our essence resides'' (p 251). Due to slow thought we have an
inexhaustible potential for creativity over which we dispose freely and with
responsibility at the same time. Free will, responsibility, and human dignity
thus belong together inextricably. The resulting paradox - a potentially free
mind vis à vis deterministic physical laws - may be resolved if ''bridge laws''
between mental and physical truth are found (p 253). Until proven otherwise,
free will and personal responsibility should be assumed.

In chapter 13 ''False false belief belief and true false belief belief'' Roeper
explores the interface of language and ''theory of mind'' (ToM). ToM is the
capability to entertain intentional attitudes such as knowing, believing,
desiring, etc., about propositional content. Specifically, ToM is expressed by
the awareness that we and others may have false beliefs (FB) about propositions.
The representation of propositions is crucially linked with language since
syntactic devices such as subordination may be needed to express the intention.
For example, the epistemic verb ''think'' is linked with a clausal proposition
''think that p'' (p 257). Failure to pass a FB task may be interpreted as having a
cognitive problem with 'other minds' or as having correct FB (implicitly) but
having a linguistic problem with subordination or not knowing how to map the FB
onto the proper linguistic structure. If we want to express the embedding of one
thought into another on the conceptual level, this is quite hard to achieve
without overt recursive syntactic devices such as subordination. In the
juxtaposed sentences ''John told a lie. The statue of liberty was turned upside
down.'' (p 265), the link between the two must be mentally construed in the
absence of any overt subordinate conjunction such as ''that''. The function of
language is to support the externalization of conceptual recursive relations
onto a mind-internal language-conceptual interface: ''Language, then, is like a
mental blackboard. It enables us to do operations inside operations...'' (p 266).
We carry out syllogistic reasoning over propositions stated on this mental
blackboard. If a child fails in a FB task, there might be a linguistic
computational problem but the concept may be there, albeit in a different and/or
weaker representational format. Roeper argues that with a modular perspective we
are less prone to underestimate the child's cognitive capabilities and may
prevent negative consequences for her.

In chapter 14 ''The idea behind the concept of 'idea''' Roeper discusses the
relation between language and mind. He argues that our representions of reality
are distorted, however, paradoxically, these distortions allow us to cope with
reality. The mind selects those aspects that are vital for its purposes,
according to its species-specific inherited biology. On this background, when we
are to judge whether a child has grasped a concept or not we have to take into
consideration the overall modular makeup of the mind. Within each module, the
reality may be represented in different ways and not every module may represent
reality target-like (yet) or the translation between the modules may be impaired
(p 285). Indeed: ''We cannot be fair to one another without accepting the modular
independence of many aspects of mind.'' (p 286).

On the relation between ideas in the human mind and principles in the physical
world, Roeper proposes that physical objects (e.g., arms and legs) contain
ideas (e.g., of movement) so that if I break my arm I haven't lost my idea of an
arm, indeed, but the physical idea of movement is interrupted (since I cannot
move my arm anymore). While the mechanism (e.g., movement of limbs) as such is
unintentional its use is highly intentional. That's why we cannot move from the
intentions (e.g., communication) to the automatic mechanism by which it is
acomplished (grammar). In the same sense, there is no (intentional) language
learning but only language growth in a strong mentalist perspective.

In the last chapter, 15, ''In defense of dignity'' Roeper tries to answer the key
question ''How can we develop an ethical philosophy that uses our incomplete
knowledge in a responsible way?'' (p 293). He claims that an answer to this
question requires ''a new ethical culture'' (p 293). In science this ethical
culture is expressed in ''humility before the complexity of human nature'' (p 295)
and respect for human wisdom alongside scientific wisdom. Such a humanitarian
science would express in its theories our ''deepest intuition about human
nature'', namely ''that all people should respect each other's dignity, and
especially each child's dignity.'' (p 296). Over-simplifying cognitive models
threaten dignity, whereas the modular theory in the sense of a ''full theory'' of
the mind guarantees it: ''A full theory will leave our sense of personal dignity
intact'' (p 297). Therefore, he speaks of ''ethical modularity''. Had this
principle of good scientific conduct been applied in earlier times already,
misguided developments in the area of cognitive science could have been avoided.
He suspects any science in which humanistic insight is not compatible with
scientific ones (p 299). Thus, our notion of ''free will'' should not easily be
given up since it opens the floodgates to control and manipulation of children.
On the contrary, the notion of creativity, which is inherent in generative
linguistics, corroborates the idea of a free will.

The only challenge he sees in the modular theory of mind with respect to dignity
is the threat that in a modular mind the sense of integrity might be missing. He
is also concerned whether the mechanistic view in linguistics, as witnessed by
the use of computer analogies, fits with integrity, free will, and creativity (p
300). However, ''slow thought'' in the service of the self may counteract this
threat. It binds together information from various domains, thus creating an
overarching ''frame'' (p 301) that may integrate our actions.

Roeper has written a highly inspiring book for scholars and for interested
laymen alike. It is the distillate of his lifelong academic studies in
generative language acquisition research. In its linguistic substance the book
is profound, rich, and detailed. It partly reproduces known facts but it
provides also novel findings on how young children acquire a wide range of
linguistic structures, starting from first words and ending with a fully
recursive complex grammar.

In the following, I will evaluate in more detail the exploration-based method
and the ''ethical modularity'' claim.

Throughout the empirical chapters, mostly in parts II and III, Roeper proposes
many linguistic explorations which parents (and other persons) may apply in
informal play situations with their children. These exploration have the
three-fold goal of (i) enabling parents to conduct small-scale linguistic
research on their children themselves, (ii) informing them where in the course
of language acquisition their children are, and (iii) revealing the working of
the mind. As for the first goal, Roeper's intention is ''to bring grammar to life
for everyone'' (p 6). The explorations therefore have an emancipatory and
educational effect on parents who do not have to remain awe-struck in front of
the increasingly high-tech methods by which cognitive and language development
of infants and young children are studied nowadays, especially neuroscientific
methods. As for the second goal, the children's reactions show how they make
sense of the problem on various points of the developmental trajectory. Since
the little tasks do not have a ''correct'' solution the child cannot fail in them.
Even a reader who does not have ready access to children as informants may
profit from the explorations. They function like illustrations to findings in
the literature or to claims by linguists on how a particular linguistic
phenomenon, e.g., plural, is acquired. As for the third goal, the explorations
are particularly suited for studying subtle grammatical and semantic
differences. Generally, Roper appreciates information from any area: ''All
evidence should be taken seriously...'' (p 6); at the same time, ''all data should
be treated with both respect and suspicion, whether they are anecdotal or
statistically robust'' (p 12). The most telling evidence with the greatest gain
in knowledge may actually come from unusual data, unexpected rare sentences
which suddenly illuminate the researcher's view on a certain aspect of
cognition: ''..., it is the extremes that provide the most insight.'' (p 7).

The most controversial claim in this book concerns ''ethical modularity'', namely
that the modular theory of mind may preserve the dignity of a child. First of
all, it should be beyond doubt that the dignity of a child should always be
preserved: a child having failed in a particular language (or cognitive) task
should incur no disadvantages such as demotion to a lower class or to a class
for handicapped pupils, misclassification as imbecile, or personal offence.
However, it may not be that easy to achieve this goal. Can cognitive science
help here? Roeper is convinced that by maintaining a modular theory of mind we
can avoid harm to the children that are entrusted to our care. The modular
theory may lend itself less easily to misuse than alternative theories. The
argument claims that a modular theory naturally preserves the dignity of the
child because the failure of a child in one task (if a failure at all) will
remain a local one in just one module whereas all other modules may still be
intact. The acknowledgement of the child's general cognitive integrity preserves
her dignity. In contrast, a theory that generalizes a local deficit to a global
one is more likely to lead to a violation of the child's dignity with all its
negative consequences.

Before evaluating Roeper's main concern - the ethical implications of our
scientific conduct - I will first analyze the logic of his argument. I believe
that the preservation of dignity does not logically follow from entertaining a
modular theory of mind. More generally, the content of a theory in and of itself
cannot preserve dignity. The theory belongs to the scientific domain and dignity
belongs to the ethical domain. These two domains are incommensurate and should
not be made commensurate at any price. By moving from modularity to dignity in
one step the scientist crosses the border between theory and ethics unnoticed,
thus violating a necessary precondition for preserving dignity. This
precondition is to be capable of preserving or damaging it in the first place
which only conscious subjects can do but which does not follow from the
scientific theory. Now the scientist who maintains the modular theory is such a
conscious subject. She may have found a workable solution for the problem of how
to preserve dignity by showing that dignity may follow from modularity - no
matter if that explanation is logically true. This analysis does not exclude
that other scientists who advocate other theories may not also find such a
workable solution in their scientific and ethical practice. Whether scientists
succeed or not in this endeavor depends on which aspect of their theory they
bring to bear on their behavior. Depending on this choice quite different
outcomes may ensue. In fact, having a modular theory or not may equally well
result in preserving or damaging dignity. Modularity has other aspects which may
threaten dignity. Actually, it is Roeper himself who adduces the best
counter-argument against modularity as an ethical theory: ''I think efforts to
subdivide human abilities lead to primitive definitions of human nature that
inevitably damage our sense of mutual regard.'' (p 296). An anti-modularist who
imitates Roeper's reasoning couldn't have phrased it any better. In a recent
weblog, Roeper critically addresses potential negative consequences of a modular
theory in foreign policy and how they can be overcome

As there are good and bad stories about a modular theory there are good and bad
stories about other theories as well. Take Piaget's constructivism which Roeper
criticizes for unduly overgeneralizing a local linguistic deficit to a general
cognitive deficit. Suppose it was true that a language deficit is just a sign of
a general cognitive deficit. Would that automatically lead to an unjust
treatment of the child? Not necessarily. Other aspects of constructivism may
guard against it such as the idea that the child actively construes knowledge by
interacting with the environment. Thus, a constructivist might try to foster
this process therapeutically and thereby remedy the assumed general deficit.
These examples show that different ethical consequences may follow from the same
theory as well as the same ethical consequences may follow from different
theories. What can we conclude from this? Entertaining a certain kind of theory
neither exempts a researcher from constant monitoring of her proper ethical
behavior nor does it condemn her to inevitable failure in this respect.
Furthermore, granting good moral conduct to one's scientific opponents is part
of the general striving for preserving dignity. It is, of course, no excuse for
looking away when obvious misconduct occurs anywhere. By emphasizing the ethical
implications of scientific theories the author makes us aware of the need for
the constant reflection of ethical issues that are inseparably connected to our
scientific research. This is Roeper's great merit. He is rightly concerned with
how we can behave ethically properly in a situation of imperfect knowledge. In
this respect we may recall Kant's (1800) famous three questions: What can I
know? What ought I to do? What may I hope? The first question ''What can I know?''
relates to our scientific striving for knowledge. In the context of the present
book it relates to finding out how the human mind works in general and how
children acquire language in particular. Roeper has come a long way on this path
of knowledge. The second question ''What ought I to do?'' relates to the proper
ethical conduct and moral responsibility that comes as an obligation with our
scientific research. Roeper's answer is to be cautious and not overstate
theories, to stress the creative nature of human cognition, and to be explicit
on potential negative consequences of our theories. Thanks to Roeper's
insistence on this second question this book may instigate a necessary
discussion on our responsibility towards the subjects of our scientific study,
not only in the area of language acquisition. The third question ''What may I
hope?'' invites us to speculate on the relation between scientific and moral
issues in the future. Here, Roeper's expressed hope is that we arrive at a
better integration of both. While our intellectual knowledge has greatly
expanded within the relatively short history of scientific research our ethics
does not seem to have so at an equal rate. However, science and ethics belong
together, maybe not exactly in the way Roeper laid out in this book, but in a
philosophical frame whose exact scope still has to be determined. Such a
discussion is well in the spirit of enlightenment.

Taking Kant's three questions together, they culminate in a fourth one ''What is
a human being?'', which is the deepest humanistic question. Roeper has vigorously
posed this question in this book. We all are invited to answer it through our
research and in our life.

Kant, Immanuel (1800/1992). _Lectures on logic. The Cambridge Edition of the
works of Immanuel Kant_. Cambridge: Cambridge University Press.

Roeper, Tom (1999). Universal bilingualism. _Bilingualism: Language and
Cognition_ 2(3), 169-186.

Annette Hohenberger is Assistant Professor for Cognitive Science at Middle East
Technical University, Ankara, Turkey. Her research areas are mainly
psycholinguistic: development of language, cognition, and action; sign language;
language production; language and memory.

I would like to thank Tom Roeper for his helpful comments on an earlier draft of
this review and for making available to me additional information on the topic
which helped me refine my evaluation. I would also like to thank Samet Bagce and
Emre Keskin for their helpful comments.
- A.H.

Format: Hardback
ISBN: 0262182521
ISBN-13: 9780262182522
Pages: 355
Prices: U.S. $ 37.95