Juan Uriagereka, (1998), Rhyme and Reason: An Introduction to
Minimalist Syntax, MIT Press, Cambridge, Massachussetts.
694p, 94.50 USD.
Reviewed by Ahmad Lotfi, Esfahan Azad University
"This book takes the form of a dialogue between a linguist
and another scientist. The dialogue takes place over six
days, with each day devoted to a particular topic ... "
(from the blurb for Rhyme and Reason)
While walking along the gray waters of a river in a magnifi-
cant afternoon, the LINGUIST Reviewer noticed someone else's
"What's up?" asked the shadow.
"Nothing, my fellow LINGUIST Subscriber, nothing. Just a
review of a new introduction to minimalist syntax."
"Oh! You don't mean that Rhyme and Reason by Juan Uriage-
reka, do you?"
"So you've heard of it! ... That's the very book we going
"Mr. Such-and-such, why don't you just write your review as
others do? Why should I indulge in a boring discourse with you?"
The Reviewer swallowed his pride and said, "Patience. Perhaps
our dialogue is not as smart as Uriagereka's. But it still
gives you both a glimpse of the ideas covered there, and a
taste of the atmosphere of the book, too. ... An unusual
review for an unusal book. fair enough, isn't it?"
And so it was that the two began the dialogue you are about
(The LINGUIST) S(ubscriber): I feel Rhyme and Reason must be
good fun! By the way, why should the dialogue between the
linguist and the other take place over six days?
(The LINGUIST) R(eviewer): Who knows? Perhaps it's an allusion
to Genesis, the First Book of Moses! ... God did this and that,
and it was the first day. ... Rhyme and Reason is not to be the
Bible of the linguist, but it does give an account of the
genesis of minimalist syntax.
S: The what of minimalist syntax?
R: Look! I've promised Andrew to keep my review under 2000
words. Perhaps if you ... .
S: All right, all right. Just proceed with your synopsis of
the book, then. It doesn't bother you if I interrupt from
time to time to make comments, does it?
R: Not at all. It'll be then some evaluation of the book the
On the first day of their encounter (Chapter One: The
Minimalist Viewpoint), Plato's problem of acquisition of knowl-
edge in the absence of the relevant experience is discussed in
reference to language acquisition. As the linguistic stimulus
for the acquisition of such knowledge is too poor, the linguist
concludes that the core of linguistic knowledge must be innate.
Language, in this sense, is some kind of biological growth re-
sulting in the ability to use finite means in order to produce
an infinite number of sentences. Universal Grammar (UG) is that
theory of language which views human language as a procedure for
generating such structures under such circumstances. Chomsky's
principles-and-parameters approach to the question of innate
knowledge assumes that universal natural laws of language
(principles) define the framework within which UG-bound
variations (parameters) are permitted. In the absence of the
relevant negative evidence on the impossibility of certain
structures in a language children have to rely on direct positive
evidence (actual utterances) in order to set the parameters of UG.
Language changes are then viewed as children's acquisition of
parametric values different from their parents' but still
confined to the restrictions of the acquisition device. While
still loyal to such an approach, the Minimalist Program tries
to translate this theory of syntax into more abstract terms of
conceptual necessity and optimal economy.
S: This naturalistic approach to language and language
acquisition must urge the linguist, then, to think of some
evolutionary pressure behind such innate requirements. But
what can natural selection have to do with such design
requirements of language?
R: To Uriagereka, actually nothing. Language is an exaptation
with its elegant aspects not being evolved. Anyway, even if
this view of language contradicts its own fundamental claims
to the biological basis of language, such criticisms should be
leveled at the program itself rather than at Uriagereka's
introduction to it.
S: What a day! The second day must be concerned with the details
of the program itself, I suppose.
R: Exactly. Chapter 2 (Notation and Reality) focuses on
Phonetic Form (PF) and Logical Form (LF), the only levels of
linguistic representation where linguistic competence interacts
with performance systems of articulatory-perceptual and
intentional-conceptual matrices respectively. Computational
System for Human Language (CHL) selects different matrices of
features, i.e. lexical/functional items, from the lexicon to be
introduced into the syntactic derivations. Such derivations
converge at the interface levels of PF and LF if the external
conditions of these levels are met. Otherwise, they will crash.
The satisfaction of these conditions is referred to as the
Principle of Full Interpretation. Furthermore, the convergent
derivations will be subject to some Principle of Optimality
according to which the most economical convergent derivations
will block the other convergent but less economical ones.
Finally, morphological strength of lexical items plays an
important role in the initiation of syntactic movements as the
checking requirement of strong features necessitates such
structural dislocations. ... My dear friend, how am I to
interpret that smile of yours?
S: Well, I was just thinking of the modifications of the
original MP model that Chomsky introduced in his Minimalist
Inquiries (1998). For instance, he introduced some new terms
and concepts instead of the more familiar ones recurrent in
MP. I mean such terms as EXP= <PHON, SEM>, EPP-features
(perhaps instead of strong features), dislocation (for
movement), Agree, phase, ... . Or just think of how Chomsky
considers his own distinction between 'delete' and 'erase'
dubious (MI, p. 55). Uriagereka couldn't keep abreast with such
modifications in his work, could he?
R: Naturally not. But I don't still understand what you're
S: Well, ... perhaps Chomsky's ideas are too slippery to be
daringly worked out into introductory works or textbooks. It
seems a bit hopeless to put such a work as Uriagereka's together,
and, while the ink on the paper is still wet, find it already at
the edge of being outdated! It's too early for Uriagereka's to
join Radford's (1988) and Haegeman's on the dusty top shelf,
R: Mmmm ... . Perhaps you have got a point there. Anyway, no one,
at least not Uriagereka, should be blamed for that!
"The Third Day: Phrases and Linearity" is the chapter where Uria-
gereka elaborates on the differences between externally determined
and conceptually necessary conditions. The first type are those
whose presence is due to the requirements of the subsystems of
grammar other than CHL itself, i.e. those functional at interface
levels. Conceptually necessary conditions, on the other hand, are
necessary due to some property of CHL itself. The operation Merge
and the property of projection belong to the former and latter types
respectively. Whether it is this or that type of condition cannot
be decided a priori. Case is a good 'case' in that it may be due to
the LF requirements of certain elements, PF linearization, or some
other properties of the system.
Linearization --the process of mapping the internal hierarchy
of phrases onto PF linear objects, and Merge are behind the property
of binary branching: the structural relation of command, which is
derived from Merge, results in some structural asymmetry between an
element higher in the hierarchy and those dominated by its mother.
This asymmetry is later mapped into PF timing sequence (precedence).
It is then predicted that specifiers precede heads, and complements
follow them even if in certain languages such structural require-
ments are SUPERFICIALLY violated.
S: I hope it is a merely STRUCTURAL requirement. It will be then
pure accident that generativist studies have been primarily con-
cerned with the English language. To tell you the truth, I sometimes
feel English is becoming a modern Latin language with a grammar
so 'THE-Theory-friendly' that may finally turn into a 'universal'
model for all other human languages. If they fit in, then all right.
Otherwise, the differences may be "superficial"-ized.
R:Even if one accepts your point (that I'm not sure I can), it is
still irrelevant to our talk this evening. You systematically
forget that assessing an introduction to a theory is different
from assessing the theory itself.
S: And you systematically forget that an introduction to a theory
cannot be better than the theory itself. It inherits both the
merits and demerits of the theory. ... All right. Let's just
call the chapter a day, and get along with the rest of the story
R: Cyclic Transformations are the major topic of discussion on the
Fourth Day. The most important operations covered in Chapter 4
are MERGE, associating two independent syntactic objects in order
to form another SO, MOVE, associating two SOs so that one dominates
the other, ADJOIN, associating a SO to a target SO so that a new
segment of the target is produced, and COPY that makes a copy of
the moved element (connected to each other via a chain).
All human languages are essentially the same at the interface
level of LF. Due to the existence of overt movement in certain lang-
uages, it follows that some covert LF movement must be necessary,
too. While overt movement involves the movement of categories,
covert movement must necessarily be the dislocation of features
in order to satisfy the economy requirement of derivation according
to which only the necessary minimum amount of the material, i.e.
formal features, need to be dislocated unless some extra materials,
i.e. PF features, are pied-piped to the feature in question in order
to check and finally delete the strong features of the target.
The term procrastination captures the tendency of the system to
force covert movement whenever possible. The need to have overt
movement, then, is related to the existence of strong features,
i.e. those features that trigger overt movement in order to be
finally checked and deleted.
S: Uriagereka, in a sense, reduces overt movement to the morpho-
logical requirements of the target, namely its strong features.
But when it comes to the definition of a strong feature, he
restates strength as that attribute of some features that triggers
overt movement. It's like saying apples drop because they have the
property of dropping. This is hardly useful as an explanation.
R: But he does imply that this is NOT the reason for overt move-
ment. It is perhaps just the mechanism for dislocation to take
place. And that's why he considers the notion 'metaphorical'.
S: It's then mere rhetoric. One wonders what the use of such a
notion is if it does not 'explain' anything. Using Chomsky's
favorite terminology, can't one dispense with such a notion?
R: Chomsky (1998) seems to have already done it! He prefers to
invest in the term EPP-features instead, though there is no
guarantee that one term is superior to another as the thought
behind both doesn't seem to have changes a lot.
In Chains and Their Checking Domain (Chapter Five), the author
is primarily concerned with the Last Resort Condition-- Move be
possible if and only if some features (of either the moved element
or the target) are thereby checked, the Minimal Link Condition--
the landing site of movement be optimally close to the original
position of the element to be moved, and the Condition on Chain
Uniformity-- the head and the tail of the same chain be identical
in categorial content. These conditions (and economy consider-
ations of the design, to be sure) put constraints on such
transformational operations of the CHL as Attract, Affect, Delete,
and Erase. The point is that such necessary conditions as the LRC,
MLC, and CCU and the economy strategies (as some ranking criteria)
cannot impose conflicting constraints on the derivations. Finally,
the Principle of Chain Integrity is introduced according to which
it will be either the case that chains are extended by raising the
head, or eliminated by deleting the tail.
S: As far as this "movement-to-check" business is concerned,
minimalists seem to find it difficult to make their minds on what
after all triggers the movement: the features of the target, those
of the element to move, or both. We first had Chomsky's Principle
of Greed according to which constituents move only out of their own
self-interest. Then Lasnik proposes his Enlightened Self Interest
assuming some altruism on the part of the element (that can also
move in order to check the morphological requirements of the target).
And finally, Chomsky changes his mind (perhaps under the influence
of Lasnik's) as he proposes some suicidal greed . Don't you think
Uriagereka trivializes the whole issue when he merely mentions that
elements move to satisfy the morphological requirements of ONE OR
THE OTHER element?
R: Agree. Even earlier introductions to the theory, like Radford's
Syntactic Theory and the Structure of English: A minimalist
introduction (1997), gave a better coverage of the issue. All I
could find on Greed throughout Uriagereka's whole work was a
single note of eight lines on page 586, and even a shorter one
on Enlightened Self-Interest.
S: Well, this must bring us to the closing day of the dialogue
between the Linguist and the Other. I hope it's a happy ending
for our story. What happened on that day?
R: On the Sixth Day (Words and Their Internal Domain), the Linguist
and the Other focus on the nature of words. Along with linguistic
variation, strength, uninterpretability of some features, and
movement itself, words are viewed as some imperfection to the
elegance of the language design, some departure from optimality in
the system. Uriagereka suspects that these imperfections are not
unrelated to each other, though the nature of the relation(s), if
any, is still in shadow.
Uriagereka is also concerned with the role of words within
this system. On the one hand, the absence of many logically possible
words suggests that they are the product of syntactic processes.
On the other hand, the fact that words are not as productive,
systematic, and transparent as syntactic processes argues against
this approach to words. It is then argued that word structures, as
a product of performance and convention, are systematic to the
extent that they are affected by syntactic residues and learning
strategies. The "end of syntax" is too far away to see!
S: You don't intend your final sentence to have two readings,
R and S together: .... Ha ha ha!
S: What's your general evaluation of the book?
R: One of the best introductions I've ever read. Topics are hit
at depth. The dialectic method of presentation adds to this depth.
Good illustrations (216 ones in total), too.
Perhaps Uriagereka is more successful in introducing the theory
than in his attempt 'to reconcile generative grammar with the
punctuated equilibrium version of evolutionary theory' as the blurb
for the book suggests.
The book could be wider in its coverage of recent issues in
theoretical linguistics in order to take (a better) care of such
terms as greed, enlightened self-interest, ellipsis, ergative/
absolutive cases, mood, etc. This limitation, however, seems quite
natural as a wider range of coverage could transform the work into
a teacher's monologue addressed to a first-year student of
linguistics rather than a lively dialogue between a linguist and
another scientist. You must read it!
Chomsky, N. (1995) The Minimalist Program, Cambridge, Mass.:
Chomsky, N. (1998) Minimalist Inquiries: The Framework, MIT
Occasional Papers in Linguistics. Cambridge, Mass: MIT Press.
Haegeman, L. (1994) Introduction to Government and Binding
Theory, Oxford: Blackwell.
Lasnik, H. (1995) Case and expletives revisited: On Greed and
other human failings, Linguistic Inquiry 26, 615-633.
Radford, A. (1988) Transformational Grammar, Cambridge Univer-
Radford, A. (1997) Syntactic Theory and the Structure of English:
A minimalist introduction, Cambridge University Press.
Reviewer: Ahmad Reza Lotfi, Assitant Professor at the English
Department of Esfahan Azad University, where he teaches
Linguistics to graduate students of TESOL. His research interests
lie in minimalist syntax, second language acquisition studies in
generative grammar, and Persian linguistics.