<< . .

( 19)

. . >>

you assume . . . OK, it is just wandering through the space of possibil-
ities, it is tinkering. But, as you learn more, you may find out that it is
not true at all, maybe most of biological evolution is like the Fibonacci
series. There is a tradition in modern biology of serious scientists
who have tried to exploit that idea. The most famous one is D™Arcy
Thompson,22 who tried to show that you could account for large as-
pects of the nature of organisms by looking at biophysics, basically:
what kinds of forms could there be? Actually Goethe did something
similar.23 He had interesting ideas, some of which turned out to be
right, I mean, not the way he thought, but basically right: in plant
growth everything is a replication of the same structure over and over
again, the stem and the leaf; he kind of guessed, it™s a mixed story,
but kind of right. With D™Arcy Thompson this becomes real science.
Not much was done with it, because it probably was too hard. But that
opened a tradition. The next famous person who picked that up was

An interview on minimalism

Alan Turing.24 This is not too well known outside biology. Turing is
mostly known for his mathematics, but he also worked on biological
problems. He was a serious scientist and he was interested in showing
how, if you have a thermodynamic system of some kind and some sin-
gularity exists, a slight perturbation, it might lead to a discrete system,
suddenly. So, he was interested in things like zebra stripes: how come
zebras have stripes instead of just some mess? And he tried to construct
models in which you™d get things like zebra stripes, just out of physi-
cal processes with a tiny perturbation which changes things around.
And the mathematical models are apparently right, so I am told. The
question whether it works with zebras is another problem, I think the
current belief (I am no expert) is that for zebras it probably doesn™t
work, but for angelfish it probably does work. There™s a certain kind
of fish that has some weird stripes all over the place and apparently
the Turing models or some modification of them do a reasonable job
for explaining that.
At the level of very simple systems, a lot of this is pretty much
assumed. Mitosis is a case in point; nobody thinks there are genes that
tell the breaking cell to turn into spheres, just as you do not have a
gene to tell you to fall if you walk off the roof of a building. That would
be crazy, you just fall because physical laws are operating, and it is
probably physical laws that are telling the cells to break up into two
spheres. Well, another case that is generally assumed is the shell of
viruses, which are polyhedrons and in fact icosahedrons. It turns out
just by pure geometry that there are only certain kinds of forms that
can appear and be stable and fit together. The viruses pick one of those
forms and they pick the one of the possible geometrical figures that is
closer to a sphere, so they don™t pick pyramids, they pick icosahedrons.
Maybe that involves selection, but the possible viral shells are assumed
to be determined just by physical law. Or take the honeycomb of bees,

On nature and language

which is again based on polyhedrons. There are other things: there is
an organism “ nobody even knows whether to call it an organism “
called a slime mold, which begins with little organisms and they all
hang together and then become a bigger organism, and finally they
split up and become separate organisms. This happens in a regular
fashion and I understand that the mathematics of this is pretty well
worked out. There is some fairly straightforward physical property that
will lead to this complicated-looking behavior once it is operative. So,
superficially, that might look like tinkering and fitting some environ-
ment, but in fact it is probably just some slight change that led to
this happening. How far does that go? Most things are just not under-
stood, so you don™t know how far it goes. When you go beyond simple
structures, you are guessing what might have happened, and when
something is learned, the guess often turns out to be wrong, because
you just can™t guess, there are too many possibilities, many not yet even
imagined. The evolution of the eye, for example, has been extensively
studied, and a standard conclusion was that it had evolved indepen-
dently about fifty times. Recent work has found that there is a single
origin, and a single “master control gene” for all eyes in the organic
world.25 Then, over billions of years, evolutionary processes (natural
selection functioning within a structured “channel”) gave rise to many
kinds of eyes, superficially very different, but with deep uniformities.
Now let™s turn to language. It appears to be a fact that language
is biologically isolated. Let™s look again at Hauser, which is the ency-
clopedic study of evolution of communication, a study of compara-
tive communication, really. Language doesn™t even fit in his taxonomy.
Human language is the exciting topic, so the book starts with language,
it ends with language, and in between there™s comparative communi-
cation studies. But in it there is a taxonomy of possible systems and
language does not belong. The possible systems include non-human

An interview on minimalism

primate calls, bird songs, etc. There are systems related to survival,
mating, and reproduction, and there are those involved in the identifi-
cation of the caller and so on. That™s about it. Language doesn™t fit in.
You can use language to identify yourself, for reproduction, for warn-
ing about predators. But one can™t study language seriously in these
terms. Language simply has no place in the taxonomy. In fact, Hauser
kind of mentions this, but without making clear the consequences of
what he is saying. He says that everything in his book is “irrelevant to
the formal study of language”; well, “irrelevant” is too strong, but that
is his statement. But what is the formal study of language? The answer
is: virtually everything about language. He may have in mind rules in
some notation, but it isn™t that: “the formal study of language” in-
cludes all the work that seeks to determine the nature of language,
just as “the formal study of bee dance” includes virtually the entire
literature on the topic. So, whether it is syntax or semantics or phonol-
ogy or pragmatics or whatever you call it, that™s the formal study of
language. If everything in the book is “irrelevant to the formal study of
language,” it is just another way of saying that language doesn™t belong
in this taxonomy. And apparently that™s true. He is certainly trying to
make a serious effort to show that language belongs, but when you
look, it turns out that it doesn™t fit, whether we have in mind the prop-
erties of language or its various “functions.” When Hauser gets to
the last chapter of the book, called “Future Directions,” he speculates
about how we might some day be able to say something about the evo-
lution of these systems, because now we can say essentially nothing.
In the case of language, what he says is: “look, there are two problems;
you obviously have to memorize a lot of words and you have to have a
generative system, which is going to give you an infinite array of expres-
sions, so something has to deal with those.” Well, how do you do it? The
infinite array of expressions, he just drops: he mentions the problem

On nature and language

with no speculation, which makes sense, because there is no serious
speculation. What about the explosive growth of the array of words?
He observes that there is little to say about this, either. It is not like
animal calls. Word learning, he points out, must involve a capacity for
imitation; so humans have an innate capacity for imitation. Of course,
far more than that, as he recognizes. What about the capacity for imi-
tation, then? Well, that turns out to be a total mystery too. According to
Hauser, that is not found in any relevant form elsewhere in the organic
world, and there™sno way of knowing how that came about, so he (in ef-
fect) concludes. So, it™s a total dead end. There is essentially nothing to
say, language is off the chart. That is the basic conclusion that follows
from his comprehensive review of comparative communication.
That doesn™t mean that language is not the result of biological
evolution, of course we all assume it is. But what kind of result of bio-
logical evolution? Well, here you have to look at the little bit we know.
We can make up a lot of stories. It is quite easy: for example, take
language as it is, break it up into fifty different things (syllable, word,
putting things together, phrases and so on) and say: “OK, I have the
story: there was a mutation that gave syllables, there was another mu-
tation that gave words, another one that gave phrases . . . another that
(miraculously) yields the recursive property (actually, all the mutations
are left as miracles).” OK, maybe, or maybe something totally different;
the stories are free and, interestingly, they are for the most part inde-
pendent of what the language is. So if it turns out that language has a
head parameter, same story; if it doesn™t have a head parameter, same
story. The story you choose is independent of the facts, pretty much.
And that™s going to be the case until you know something. You can
make up stories about the eye, about wings and so on. What happened
is what happened, it is not necessarily the story you chose. And look-
ing at the marvelous adaptation of some system to its environment,

An interview on minimalism

when that is what we find, just sets the problem, it is not the answer,
contrary to common misunderstanding.
Going back to language, what you have is a system that is, as
far as we know, essentially uniform. Maybe there was some speciation
at one point but only one species survived, namely us; there seems to
be no variation in the species. True, we find Williams™ syndrome and
Specific Language Impairment. But that™s not variation in the species
in any meaningful sense: those are deviations from the fixed system
that occur now and then, but the basic system seems to be uniform.
In other words, kids learn any language anywhere, as far as we know,
which means the basic system is uniform. Nobody has found any ge-
netic differences; maybe there are some, but they are apparently so
slight that we can™t detect them. So, it is a fundamentally uniform
system, which means that since its emergence there has not been any
significant evolution. It has just stayed that way. People have scattered,
there are groups of people that have been separated for a long period,
but nobody can detect any language difference. So it™s apparently a
recent thing, too recent to have undergone much evolution. There is
also a point that Jerry Fodor has recently stressed:26 language is dif-
ferent from most other biological systems, including some cognitive
systems, in that the physical, external constraints that it has to meet
are extremely weak. So, there™s some innate system of object recogni-
tion: infants can identify object constancies; they know things don™t
go through barriers, etc. But that system, whatever it is, has to be at-
tuned to the outside world; if you had a system that had objects going
through barriers and so on, you couldn™t get along in the world. So that
system is sort of controlled by the outside world. Then it makes sense to
speculate that it was selected “ this is a speculation, but plausible, like
echolocation. On the other hand, language doesn™t have to meet that
condition, or it has to meet it in an extremely weak way. You have to be

On nature and language

able to talk about the world, somehow, but there™s any number of ways
of doing that. The fundamental condition that language has to meet
is that it can be used, that the person who has it can use it. Actually
you can use language even if you are the only person in the universe
with language, and in fact it would even have adaptive advantage. If
one person suddenly got the language faculty, that person would have
great advantages; the person could think, could articulate to itself its
thoughts, could plan, could sharpen, and develop thinking as we do
in inner speech, which has a big effect on our lives. Inner speech is
most of speech. Almost all the use of language is to oneself, and it can
be useful for all kinds of purposes (it can also be harmful, as we all
know): figure out what you are going to do, plan, clarify your thoughts,
whatever. So if one organism just happens to gain a language capacity,
it might have reproductive advantages, enormous ones. And if it hap-
pened to proliferate in a further generation, they all would have it. In
a larger group all that is necessary is that it be shared. The connection
to the outside world is extremely weak and therefore it could be very
stable, because there is no point in changing it; there™s no advantage to
any change that takes place, or it could be stable because it just didn™t
have enough time. One way or another, it has evidently been stable.
What happened pre emergence? That™s anybody™s guess; it
seems to be absurd to regard it as an offshoot of non-human primate
calls. Language doesn™t share any interesting properties with them.
Or with gestural systems; or anything that we know about; so, you
are stuck. Language has highly unusual properties: discrete infinity is
unusual, displaced reference is unusual, the most elementary struc-
tural and semantic properties seem unusual. It is possible that what
happened is what Richard Lewontin and others have speculated:27 the
brain was exploding for a million years; it was getting way bigger than
among other surviving primates, and at some stage (for all we know

An interview on minimalism

about 100,000 years ago) some slight change may have taken place
and the brain was reorganized to incorporate a language faculty.
Maybe. That would be like angelfish stripes, polyhedral shells, etc.
The understanding of the physical channel for natural selection is so
limited that you really cannot have an opinion on this. You can make
fun of it, if you like, or you can wave a banner about it. But that doesn™t
make much sense. It is simply not understood how the physical chan-
nel constrains and controls the process of selection, beyond simple
cases. Lewontin is one who thinks that we™ll never know the answer for
human higher mental processes “ that by any method we can imagine
now, there is no way to find the answer, not just for language but for
cognition altogether. Others feel that they can do something. But
telling stories is not very instructive. You can tell stories about in-
sect wings, but it remains to discover how they evolved “ perhaps from
protuberances that functioned as thermoregulators, according to one
account. A famous case is giraffes™ necks, that was the one case that
was always referred to as the obvious example of natural selection with
a clear function; giraffes get a little bit longer neck to reach the higher
fruits, and they have offspring and so giraffes have long necks. It was
recently discovered that this is apparently false. Giraffes don™t use their
necks for high feeding, end of that story. You have to figure out some
other story: maybe sexual display like a peacock tail or some other
story, but the point is that the story doesn™t matter. You can tell very
plausible stories in all sorts of cases but the truth is what it is. You can
tell stories about the planets, as the Greeks did, in fact: nice stories, but
things don™twork that way. In the case of language, we know that some-
thing emerged in an evolutionary process and there is no indication of
any evolutionary change since it emerged. It emerged once, as far as
we know, very recently. There is no real evidence for use of language
prior to maybe 50,000 years ago or so. But the neuroanatomy seems

On nature and language

to have been in place before that, so maybe 150,000 years ago. Anyway
it™s recent. The emergence seems to be fairly sudden, in evolutionary
terms, in an organism with a very large brain, which was developed for
whatever reason, and conceivably through some reconstruction of the
brain that brought into play physical processes that led to something
that works close to optimally, like a virus shell. If the minimalist the-
sis actually gains some significant credibility, that would be not an
unreasonable conclusion; of course you have to establish the thesis.
ab & lr: So, language could have come to existence suddenly,
through a single mutation, basically in its modern form, and nat-
ural selection wouldn™t have had time to act on it. How can we
substantiate this “evolutionary fable,” as you call it in “Minimalist
Inquiries”? What kind of evidence do we have of the recency of
human language?
nc: Well, one thing is that there just weren™t a lot of humans
around, as far as anybody knows. Current estimates of the number
of individuals, I can™t reconstruct reliably from memory, but it may
have been something like maybe 20,000 about a hundred thousand
years ago, in fact a very small population, which then scattered widely.
Unlike other large organisms, humans had escaped any limited eco-
logical niche, so they were all over the place, presumably from a single
origin. They were adapted to many environments. That means very
small groups and not many of them. And then there was an increase;
I mean, nothing like the explosion of the last couple of hundred years,
but there was a substantial increase and that coincided roughly with
the appearance of symbolic manifestations, various ceremonies and
people buried with their tools, lots of things that indicate that there
was complicated social organization. That™s pretty hard to imagine
without language. So that™s the kind of evidence available. There is
also some physiological evidence: Philip Lieberman has argued that

An interview on minimalism

their larynx sank.28 Some scientists agree, some don™t. Whatever it
means, it is peripheral. On the perceptual side, there doesn™t seem
to be anything much detectable, and of course, as for thought sys-
tems, there are no records and not a great deal to learn from surviving
non-human primates, so it appears.

Scope and perspectives

ab & lr: In a recent lecture at the Scuola Normale of Pisa you
quoted English eighteenth-century chemist Joseph Black stressing
the importance, for his discipline, of establishing a “body of doctrine”
on the model of Newtonian physics. Generative grammar and, more
specifically, the Principles and Parameters framework has certainly
permitted many subtle and surprising discoveries over a broad domain,
and one may argue that a significant “body of doctrine” on different as-
pects of human language has been established. Taking for granted the
obvious fact that nothing is definitively acquired in empirical science,
what are those aspects that you would consider “established results”
in our field?
nc: My own view is that almost everything is subject to question,
especially if you look at it from a minimalist perspective; about every-
thing you look at, the question is: why is it there? So, if you had asked
me ten years ago, I would have said government is a unifying concept,
X-bar theory is a unifying concept, the head parameter is an obvi-
ous parameter, ECP, etc., but now none of these looks obvious. X-bar
theory, I think, is probably wrong, government maybe does not exist.
If Kayne is correct, the right parameterization is not the head param-
eter, but some other kinds of parameters about optional movements,
certainly plausible, possible. We just have to see. But I don™t think that
is so unusual. If you look at the history of the sciences, this is just

On nature and language

the usual situation. Even in the advanced sciences almost everything
is questionable. What I learned in college, let™s say, in science courses,
a lot of it would not be taught today. In fact, what was taught twenty
years ago would be taught differently today in physics or chemistry.
Some things are relatively stable. The Periodic Table is still there, but
elementary particles are nothing like what we were taught. In fact in any
live discipline you really don™t expect the body of doctrine to be terribly
stable. You™ll get new perspectives, things will be reinterpreted. The
changes often may not look very great from the outside but in a sense
you can say the same thing about generative grammar for fifty years.
From the outside it looks more or less the same, but from the inside
you can see that it is very different and I suspect that that will continue.
What are island conditions, for example? This has been a core topic
of research for forty years now; I still don™t think we understand that.
There™s certainly plenty of data that aren™t understood; Paul Postal29
has a recent book about it and I am sure that it has tons of data that
don™t work in any imaginable way. Such problems abound. And also,
to my knowledge at least, there is no really principled account of many
island conditions. On the other hand, something will remain stable.
The difference between weak and strong islands looks stable; maybe
we don™t understand it, but there™s something there that is stable. Also
conditions on locality and successive cyclic movement look stable to
me, at some level of abstraction. I strongly suspect that the difference
between interpretable and uninterpretable features will turn out to
be stable, though it is a recent observation, five years ago there was
no discussion about it. In some fashion metrical theory will remain.
Argument structure will also remain, as will properties of scope and
reconstruction and the recent discoveries about fine structure. The
essence of binding theory will remain, but probably will be reinter-
preted. It™s not that anything ever gets thrown out; the results about,

An interview on minimalism

say, ECP will remain but they may be parceled out in different domains,
maybe with different ways of looking at them and so on. But I don™t
feel that one can really predict much. It™s a young field, changes are
taking place fast, there are lots of unexplained things. I am sure there
are new perspectives that we haven™t thought of yet. I wouldn™t expect,
or even hope for, stability. If there™s stability, it means we are not going
to get very far because, in the stage where we are now, there are just
too many mysteries. So if the field remains stable, that means there
are going to remain mysteries. That was true for chemistry at the time
that Joseph Black wrote, the chemist you quoted, mid eighteenth cen-
tury. Let™s just consider what chemistry was like in the mid eighteenth
century and what it™s like today. Black wouldn™t be able to recognize
the current discipline. In Black™s days, it was still commonly assumed
that the basic components of matter are earth, air, fire, and water, that
water can be transmuted into earth, and so on. Chemists had a sub-
stantial “body of doctrine” at that time, they knew a lot about chemical
reactions, when they took place and how they took place, but the way
of looking at them has totally changed. Take a look at Lavoisier for ex-
ample, who founded modern chemistry and created the nomenclature
that everybody still uses “ and the nomenclature wasn™t just terminol-
ogy; it was supposed to be truth, it was designed to tell you the truth:
so oxygen is the acid generator because that™s its nature (which turns
out to be false). In one of his classifications, alongside of hydrogen
and oxygen we find “caloric,” what we call “heat.” So, everything has
changed. And he kind of anticipated it; he said at that time that prob-
ably the nature of the elements is unknowable by humans, so we can
just make some speculations. And chemistry was a pretty advanced
science by that time.

<< . .

( 19)

. . >>

Copyright Design by: Sunlight webdesign