<< . .

( 19)

. . >>

ab & lr: Sometimes speaking with specialists of other disci-
plines, people ask: what are the results of modern linguistics? Is there

On nature and language

a way of phrasing some of the results independently from the technical
language that makes them opaque for the public at large?
nc: There are things understood which you can illustrate easily,
like, say, properties of wh- movement, which are very dramatic and a
lot of them we understand at some level, for example Huang™s dis-
tinctions and island effects,30 or even more complicated things like
parasitic gaps and so on. Even very simple examples can illustrate
quite complex points. Sometimes I use examples like complex adjecti-
val constructions (English is good for this, better than other languages
with complex adjectivals), which illustrate successive cyclic movement
in the predicate phrase, even though there™s nothing visible, there is
an empty operator. But the facts are clear and you can see the same
facts that you see in wh- questions; you can state the principles that
yield the interpretive facts in John is too stubborn to talk to, that sort of
thing. There™s plenty of material like that, which is stable, easy to il-
lustrate; you can state the principles, something that is known about
the general principles. The fact that there is a component that deals
with phrase structure in some fashion and a component that deals with
dislocation in some fashion, that™s, I think, pretty clear, and also that
they have different properties, different semantic properties, different
formal properties. The same if you move to phonology. So, sure there
is a substantial body of things that can be presented in public talks;
say, anything from middle school students to college and general pub-
lic audiences. It is pretty easy to bring this kind of material to them “
and I am sure you do the same “ to get them to understand and even
see the underlying principles. So, there are many non-trivial answers.
On the other hand, if you ask for an axiomatic system, there is no such
thing, but then you can™t do it for any other science either. I mean, if
somebody asks you what are the results of biology, all you can do is
give an organized system involving natural selection, genes, Mendel™s

An interview on minimalism

results and modern genetics, and so on, and then you can illustrate
ab & lr: The Minimalist Program has led researchers to rethink
the foundations of their work, thus offering fresh perspectives on
old problems, opening new questions, etc. On the other hand, the
Program selects its own empirical domain on the basis of its stringent
criteria, thus leaving out of its scope a significant part of the previously
constituted “body of doctrine.” Is this inevitable? Do you think it is
nc: It would be nice to subject everything to a minimalist cri-
tique, but it is quite hard; because nothing resists that critique, in any
domain. So as soon as you look at anything, the best established work,
and you ask, “Can I explain this just on the basis of legibility at the in-
terface?,” the answer is no. That is true for the most elementary things,
like sound“meaning correspondence: that™s the basic data that peo-
ple use, this sound corresponds to this meaning, that is everybody™s
basic descriptive data. But that doesn™t satisfy minimalist criteria, the
stringent ones at least. A stringent minimalist criterion would say:
“The expression has to be legible on the sound side and has to be legi-
ble on the meaning side; but if it pairs up properly, that™s something to
be explained.” You are not given that datum and that would require a
much richer set of conditions imposed from the outside; in fact I don™t
think it would even be statable as a set of conditions from the outside
because in order to know that the pairing is correct, you have to know
pretty much everything. So, somehow, even that simple datum, which
every linguist for thousands of years has taken as the basic datum of
the field, isn™t available on a minimalist account. You have to try to
explain it, you have to show that the optimal solution to legibility on
the sound side and on the meaning side independently is going to give
you the right interpretation for John is easy to please, not some simpler

On nature and language

interpretation. So the thing to do, at least it seems to me, is to pick
the core pieces, like, say, phrase structure and dislocation, and ask
what components of these systems look as if they are problematic. For
instance, using the criterion that I think you had suggested earlier:
would it be in an invented symbolic system? That™s a good starting
point. If you find something that wouldn™t be in an invented symbolic
system, you have to ask why it is in language: morphology, for example,
what is it doing? And as soon as you ask, that drives you to new things,
like the difference, for example, between interpretable and uninter-
pretable features, which is quite transparent but I had never thought
about it before, at least. It never occurred to me that there is a reason for
the traditional asymmetry of agreement that we all learned in school.
If you look at it from the point of view of ten years ago, I would have
said the relation is symmetric and the traditional asymmetry is just
arbitrary convention. But it is clearly not irrational, it is an intuitive
perception of something that appears to be quite deep; the distinction
between interpretability in one position and not in another position.
So, it™s not trivial, but these things don™t occur to you until you start
asking: why is it there? But then that proceeds for everything; every-
thing that fell under ECP, under binding, under government, under
proliferation of inflectional categories, almost everything. As soon as
you begin to ask the simplest question, I think that the descriptions
that looked obvious appear quite problematic, and the questions be-
gin to proliferate as soon as you investigate. That™s true of just about
any point that you look at. In anything that you look at, you see that
the assumptions are OK at some level and in fact revealing, some are
very revealing, but then you start looking at the assumptions on which
they are based, and find that the assumptions are dubious, they are not
self-evident and sometimes not even natural. In particular, they surely
don™t follow from just the fact that the language has to be legible

An interview on minimalism

at the interface. Therefore you have to seek some other explanation
for them and either you say: “Well, I give up, explanations have to end
somewhere, it™s a mystery,” or else you look for an explanation, and
the assumptions often dissolve. Anyway, we shouldn™t accept the idea
that it™s a mystery. Maybe it is, but it™s way too early to assume that.
That is an admission of defeat that is surely premature. It could turn
out to be right, maybe it™s a mystery. We have been all along (and justly,
I don™t criticize this) willing to accept principles because they yield
results. That™s the right way to proceed, without asking why such prin-
ciples exist. However at some stage, maybe it is too early, but at some
stage it is going to be necessary to ask why the principles exist and a
minimalist approach gives one way of looking at this. Maybe there™s
some other way but I can™t think of any other way at the moment.
ab & lr: One can address the same problem of empirical cover-
age from a slightly different perspective. On the one hand, the MP relies
heavily on a theory of the interfaces, which should provide the external
constraints to be met by the language faculty. As such, MP should
promote research on the neighboring systems and the interfaces even
more than previous models. On the other hand, the program so far does
not offer much guidance for the study of systems that are assumed to
be language-related, but differently constituted from “narrow syntax,”
in your sense. Do you think this is a contingency of the current state
of research, and things could, or should change in the future?
nc: First of all, the focus on the interfaces is extremely recent;
until now, it has always been assumed, as far as I know, without any
question, that there are two interfaces. This goes back to Aristotle:
there™s a sound and a meaning, and that™s it. You look at sound“
meaning correspondences, phonetics tells you the sound, nobody
knows what tells you the meaning. That has been the general assump-
tion; and it didn™t matter much. Whether the assumption was right

On nature and language

or wrong it had no effect on the theories, because they were not de-
signed to satisfy the interface conditions. As soon as you think about
that, about the fact that the essential property of language must be that
it satisfies the interface conditions “ and that much, everybody has to
accept “ , then the question arises: what are the interfaces? It didn™t
really arise before, but now it™s going to matter. As soon as you look
at it, you see that we really don™t know.
So, let™s take the easy case: the sensorimotor interface. It has
always been assumed that there is one, but that is not in the least obvi-
ous. There might be different ones for articulation and perception, and
furthermore it is not obvious that there is one interface for either ar-
ticulation or perception. Suppose that something like Morris Halle™s
picture is correct:31 the features at some level are giving instructions to
the articulators. Well, they don™t all have to do it at the same point in the
derivation. Perhaps some give instructions at one point and then there
could be more phonological computation, then another instruction is
given, and so on. It could be a distributed system in this sense. That
is possible. I mean, why should biology be set up so that there is one
fixed point in the computation at which you have an interface? Inter-
pretation could be “on line” and cyclic, and even at each stage of the
cycle, instruction to articulators and the perceptual apparatus might
be distinct in character (rather than a single phonetic representation)
and distributed within the computation. There might also be the kinds
of interaction proposed in the motor theory of perception. These may
involve interactions between two aspects of the phonetic interface. So,
I suspect that there may well be all sorts of surprises.
On the other side, the meaning side, it seems to me that there
may be some suggestive results. A lot of the most interesting syntac-
tic work that is now being done (usually called “semantics” though it
should be considered the edge of the syntax, I think) doesn™t satisfy

An interview on minimalism

natural minimalist conditions on the language faculty: binding the-
ory, quantifier scope or even operations that appear to involve move-
ment, like Antecedent Contained Deletion. These do not easily fit in
the whole picture. For one thing the operations are countercyclic, or,
if cyclic, involve much more complex rules transferring structures to
the phonological component, and other complications to account for
lack of interaction with core syntactic rules. It is conceivable that these
are just the interpretive systems on the meaning side, the analogue
to articulatory and acoustic phonetics, what is going on right out-
side the language faculty. Nobody really has much of an idea about
the computational processes right outside the language faculty. One
could say there is a language of thought or something like that, there
are concepts, etc., but there has never been any structure to the sys-
tem outside the language faculty. Well, maybe this is the beginning of
discovery of some structure right at the edge, using operations similar
to internal operations but probably not the same. They have different
There are some interesting possibilities; for instance these op-
erations on the outside don™t iterate. So, it seems you don™t have suc-
cessive cyclic QR, successive cyclic Antecedent Contained Deletion.
That is also true apparently of the operations that probably are on the
sound side, between the internal syntax“phonology interface and the
external interface between the language faculty and the sensorimotor
system. Things that involve heaviness, let™s say, Heavy NP Shift, all the
operations that fall under Ross™s Right Roof Constraint. These also
don™t iterate. That part of the internal syntax is, in a way, peripheral.
It is not part of what one would imagine to be the essential core of
language: the mechanisms for formulating thought in internal lin-
guistic expressions. The operations of the phonological component,
broadly construed, are forced by the needs of the sensorimotor system.

On nature and language

And if these operations have properties similar to those external to
the other interface, that™s suggestive. So, maybe that is the beginning
of some kind of non-trivial study of thought systems, how they are
working at the point near the language faculty where you can gain at
least some access to them. Those are new questions, questions that im-
mediately flow from the insistence “ right or wrong “ that the internal
operations have highly systematic minimalist properties.
The general point is that “ as is normal in the sciences “ you
are trying to show how the language faculty meets certain conditions,
but you have to discover those conditions, and you expect to discover
what the conditions are in the course of the process of asking how the
language faculty satisfies them. It™s not like the case of an engineer
who is given the conditions and is told: “OK, satisfy them.” Here we
are in a process of discovery, we have to find out what the conditions
are and finding out what the conditions are is part of the process of
finding out how to satisfy them, so the two processes are going to go
hand in hand. If this whole approach turns out to make any sense as a
research topic, it should lead to a much more careful exploration of the
interfaces themselves, what™s on the other side of them. That should
be a major research endeavor, which really hasn™t had much of a place
in the subject until now.
Actually, the imaging work may be of particular interest here.
Imaging studies should be particularly valuable in sketching out the
general architecture of systems and how they interact, hence in explor-
ing the ways in which the language faculty (or the several language fac-
ulties, if that is how the picture develops) interacts with other systems
of the mind“brain. Some light is shed on these questions by “nature™s
experiments” (brain damage, etc.), but direct invasive experimenta-
tion is of course excluded. The newly emerging technologies should
provide a way to overcome some of the barriers imposed by ethical

An interview on minimalism

considerations and the diffuse effects of natural events. Even in early
exploratory stages, there are results that are quite suggestive, and it
may be possible to design experimental programs that would yield
important new kinds of information about the nature of the language
faculty and the way it is accessed and used.

Chapter 5

The secular priesthood and the
perils of democracy

The term “secular priesthood” I am borrowing from the distinguished
British philosopher and intellectual historian Isaiah Berlin. He was
referring to Communist intellectuals who defended the state religion
and the crimes of power. To be sure, not all Soviet intellectuals joined
the secular priesthood. There were the commissars, who defended and
administered power, and the dissidents, who challenged power and its
We honor the dissidents and condemn the commissars, rightly
of course. Within the Soviet tyranny, however, quite the opposite was
true “ also of course.
The distinction between “commissars” and “dissidents” traces
back to the earliest recorded history, as does the fact that, internally,
the commissars are commonly respected and privileged, and the dis-
sidents despised and often punished.
Consider the Old Testament. There is an obscure Hebrew word
that is translated as “prophet” in English (and, similarly, other Western
languages). It means something like “intellectual.” The prophets
offered critical geopolitical analysis and moral critique and counsel.
Many centuries later, they were honored; at the time, they were not

The secular priesthood and the perils of democracy

exactly welcomed. There were also “intellectuals” who were honored:
the flatterers at the courts of the kings. Centuries later, they were de-
nounced as “false prophets.” The prophets were the dissidents, the
false prophets the commissars.
There have been innumerable examples in the same era and
since. That raises a useful question for us: Are our own societies an
exception to the historical rule? I think not: they conform to the rule
rather closely. Berlin used the term “secular priesthood” to condemn
the commissar class of the official enemy; a perfectly just condemna-
tion, but normal. Another historical universal, or close to it, is that we
have a keen eye for the crimes of designated enemies and denounce
them vigorously, often with great self-righteousness. Looking in the
mirror is a little more difficult. One of the tasks of the secular priest-
hood in our societies, as elsewhere, is to protect us from that un-
pleasant experience.
George Orwell is famous for his eloquent denunciation of the
totalitarian enemy and the scandalous behavior of its secular priest-
hood, most notably perhaps in his satire Animal Farm. He also wrote
about the counterpart in free societies, in his introduction to Animal
Farm, which dealt with “literary censorship” in England. In free
England, he wrote, censorship is “largely voluntary. Unpopular ideas
can be silenced, and inconvenient facts kept dark, without any need
for any official ban.” The result is that “Anyone who challenges the
prevailing orthodoxy finds himself silenced with surprising effective-
ness.” He had only a few remarks about the methods used to achieve
this result. One is that the press is in the hands of “wealthy men who
have every motive to be dishonest on certain important topics,” and to
silence unwelcome voices. A second device is a good education, which
instills the “general tacit agreement that ˜it wouldn™t do™ to mention
that particular fact.”

On nature and language

The introduction to Animal Farm is not as well known as the
book. The reason is that it was not published. It was found in Orwell™s
papers thirty years later, and prominently published. But it remains
The fate of the book and the introduction are a symbolic il-
lustration of the general point. Their secular priesthood is bad, even
despicable; their dissidents are wholly admirable. At home, and in the
dependencies, the values are reversed. The same conditions hold for
crimes that the secular priesthood must condemn with outrage, or
suppress and justify, depending on the agent.
It is, again, all too easy to illustrate. But illustrations are mis-
leading. What is important is their overwhelming consistency, a fact
that has been extensively documented in dissident literature, where it
can easily be ignored, as Orwell pointed out in his unknown essay on
voluntary censorship in free societies.
Although this course is misleading for the reasons mentioned,
I will nevertheless illustrate the general pattern with a few current
examples. Given the consistency, contemporary examples are rarely
hard to find.
We are meeting in November 1999, a month that happens to
be the tenth anniversary of several important events. One was the
fall of the Berlin Wall, which effectively brought the Soviet system to
an end. A second was the final large-scale massacre in El Salvador,
carried out by US terrorist forces called “the army of El Salvador” “
organized, armed, and trained by the reigning superpower, which has
long controlled the region in essentially this manner. The worst atroc-
ities were carried out by elite units fresh from renewed US training,
very much like the Indonesian commandos who were responsible for
shocking atrocities in East Timor, once again, this year “ continuing
at this very moment, in fact, in camps in Indonesian West Timor. The

The secular priesthood and the perils of democracy

Indonesian killers were the beneficiaries of US training that contin-
ued right through 1998, arranged by President Clinton in violation of
the clear intent of congressional legislation. Joint military exercises
with US forces continued until a few days before the referendum of
August 30, 1999, which unleashed a new wave of army-led violence
after a year of atrocities that reached well beyond what happened prior
to the NATO bombing in Kosovo. All of this is known, but “silenced
without any official ban,” in Orwell™s words.
Let us return to the tenth anniversaries, with a few words about
each of the two examples, beginning with the atrocities in the US
dependency of El Salvador in November 1989.
Among those murdered were six leading Latin American intel-
lectuals, Jesuit priests. One of them, Father Ignacio Ellacuria, was the
rector of the major university in El Salvador. He was a well-known
writer, as were the others. We may ask, then, how the US media and
intellectual journals “ and Western intellectuals generally “ reacted to
the murder of six leading dissident intellectuals by US terrorist forces:
how they reacted at the time, or right now, on the tenth anniversary.
For today, the answer is simple. The response was silence. An
electronic search of the US media found no mention of the names of the
six murdered Jesuit intellectuals. Furthermore, virtually no American
intellectual would know their names, or would have read a word they
have written. Much the same is true in Europe, to my knowledge. In
sharp contrast, everyone can reel off the names and quote the writings
of East European dissidents, who suffered severe repression, but in
the post-Stalin period, nothing like the horrors that have been a
routine fact of life in Washington™s domains.
The contrast is revealing. It teaches us a lot about ourselves, if
we choose to learn. It illustrates well what Orwell described: voluntary
subordination to power on the part of the secular priesthood in free

On nature and language

<< . .

( 19)

. . >>

Copyright Design by: Sunlight webdesign