Universal Grammar
Slides on the net at:
Chomsky’s Epiphenomenalism
about Language
Language vs. Grammar
“Grammar” is a precise definite term while “language”
is a vague and derivative term which we could well
dispense of, without much loss.
The grammar in someone mind/brain is real while
language is not.
The aim of linguistics can be summarized by four
1. What constitutes knowledge of language?
2. How is such knowledge acquired?
3. How is such knowledge put to use?
4. What are the physical mechanisms that serve as the
material basis?
Deep vs. Surface Structure
Port Royal Grammar (1660)
It is heavily influenced by Descartes.
It aims to propose the general form of any possible
In so doing it elaborates the universal structure
underlying the “natural manner in which we express
our thoughts”.
The inner/outer aspect of language
According to Port Royal grammarians we must
distinguish between language having an inner and an
outer aspect.
Hence we distinguish between a sentence qua
expression of a thought and the physical shape of a
sentence (i.e. an utterance).
To show the structure of the mind the grammar should
reflect properties of all minds, it should be universal.
Mental Grammar
The deep structure is often only implicit and does not
get expressed. It is only represented in the mind.
The same deep structure can be realized differently in
different languages (e.g.: “Video canem currentum” and
“Je vois un chien qui court”).
The rules of this grammar are not represented in the
language user: they are simply there. Yet they must be
“learned”. But see poverty of the stimulus argument.
Transformation Rules
There are transformation rules operating from deep to
surface structure. It is the linguist’s job to figure out
these rules.
The grammarians of Port Royal are the first to
recognize the two systems of rules:
1. A base system generating deep structure.
2. A transformational system mapping these deep
structures into surface structure.
UG corresponds to the deep structure. Since it is the
expression of though, it is common to all languages.
It is thus universal. Hence Universal Grammar, UG.
The transformation rules converting the deep structure
into surface structure may differ from language to
Different outputs can correspond to the same inner
Port Royal
Within the Cartesian tradition exemplified by the
grammarians of Port Royal, the deep structure is what
constitutes the meaning (sense) in the mind.
It can be transmitted in different way (e.g.:
E.g.: different languages or different surface structures
transmit the same meaning/sense which is a mental
Nowadays UG means the initial state of a language
It is the “innate” (genetically transmitted) aspect of
grammatical rules; the language instinct (Pinker).
It is that aspect of the human mind that causes one to
learn the language.
UG qua initial state is biologically determined.
As such, it does not belong to a specific language.
UG need not be supposed to be what is universal
among languages (see Jackendoff 2002: 72ff.).
It is merely the human capacity, i.e., the initial state,
allowing one to learn a language.
The aspects of the initial sate one ends up using in
one’s learning periods depends on the stimuli/input.
Languages (inputs) affect the development of the initial
state and thus the outputs one ends up producing (cf.
switches metaphor explaining the learning of
We do not necessarily mean that it is present at birth or
in an embryo.
It rather means that it automatically appears during
the development, regardless on whether it is present at
birth or not.
It does not mean that it is free from the input of the
environment. E.g. vision capacity.
Deep vs. Surface Structure, and Creativity
The deep/surface structure distinction is what helps
explaining linguistic creativity.
The Port Royal’s distinction between deep and surface
structure implicitly contains recursive devices allowing
for infinite uses of the finite means that it disposes.
The deep structure is what gets represented in the mind
when a sentence is produced/heard (see LF).
Linguistic creativity and the argument for mental
The expressive variety of language use implies that the
brain of a linguistically competent user contains a set
of unconscious grammatical principles.
(cf. Jackendoff R. 1994. Patterns in the Mind. Basic Books
Harper Collins, New York: 6).
In adopting the language of thought hypothesis, LOT
(or Mentalese) the argument for mental grammar can be
stated along the compositional principle for thoughts, or
what Fodor characterizes as the productivity of thought.
The classical argument that mental states are complex adverts to
the productivity of the attitudes … The LOT story is, of course,
a paradigm of this sort of explanation, since it takes believing to
involve a relation to a syntactically structured object for which a
compositional semantics is assumed. (Fodor J. 1987.
Psychosemantics. MIT: 147-8)
Logical vs. Grammatical Form
Arnauld & Nicole (in Port Royal Logic 1662: 160)
highlight the difference between deep (logical) structure
and surface (grammatical) structure.
(1) Now few pastors at the present time are
ready to give their life for their flocks
the grammatical (surface) structure is affirmative, while
its underlying structure (LF) is negative.
(1) contains the implicit negative sentence (“it contains
this negation in its meaning”):
(1a) Several pastors at the present time are not ready to
give their lives for their flocks
The same with:
(2) Come see me
whose deep structure is:
(2a) I order/beg you to come see me
According to the Port Royal grammarians there is a
transformation enabling to go from (1a/2a) (deep
structure) to (1/2) (surface structure).
We have hidden underlying structure and a grammatical
transformations operating between the deep structure
(LF) and the surface (or grammatical) structure.
E.g.: the surface structure “Only the friends of God are
happy” is a transformation of the deep structure “The
friends of God are happy” and “all other who are not
friend of God are not happy”.
To understand a sentence one must grasp the sense, i.e.
the meaning (“natural order”) the speaker has in mind.
One grasps it in reconstructing its meaning, i.e. in
coming to entertain its underlying structure (LF) and
the meanings of the single words.
The fundamental principles at work are reordering and
ellipsis which enable the hearer to recover in her mind
the meaning the speaker has in her.
Linguistic explanation and
Grammaire Générale (Port Royal)
Cartesian linguistics did not confine to a mere
description of a language and its grammar.
It aimed to capture the universal (mental) structure
underlying languages.
Port Royal grammar, like modern (Chomsky’s inspired)
linguistics can be viewed as a branch of psychology or
cognitive sciences.
The general grammar is a kind of universal grammar.
As such, it differs from the special grammar which is
language specific. It differs from the grammar of
English, Chinese, etc.
Linguistics/General Grammar as a Science
General Grammar is … the rational science of the immutable
and general principle of spoken and written language, whatever
language this may be … General Grammar is a science, because
its object is rational speculation on the immutable and general
principle of language … The science of grammar is anterior to
all languages in so far as its objects presuppose only the
possibility of languages and are the same as those which guide
human reason in its intellectual operations … because they are
eternally true (Bauzé 1767).
Shortcomings of Cartesian
Linguistics (1600-1700)
The underlying assumption
UG (the abstract structure underlying a natural
language sentence) is a kind of sentence itself.
It is generally assumed that deep structure consists of
actual sentences in a simpler or more natural
The underlying assumption is gratuitous and can be
It rests on the Cartesian idea that the general principles
underlying and determining our thoughts and
perceptions must be accessible to introspection and
can be brought to consciousness with care and
If we assume that UG is unconscious we don’t have
to assume that the general principle are sentence-like
Language acquisition
Universal Conditions
They are not learned and must exist for language
knowledge to be explained.
They are the pre-requisite leading to knowledge:
principles or notions implanted in the mind … a direct gift of
Nature, a percept of natural instinct … [they] remain latent
when their corresponding objects are not present, and even
disappear and give no sign of their existence (Herder 1624).
This contrasts with the empiricist view that
our mind is a clean sheet, as though we obtained our capacity for
dealing with objects from objects themselves (Herder 1624).
The mind is not a tabula rasa.
The universal principles are innate and implicit.
Yet, we may require external stimulus to activate them
and make them available to introspection.
This is one of the main principles underlying the
psychology of Cartesian linguistics and rationalism in
general (see e.g. Leibniz).
[I]t is true that it is purely arbitrary to connect a certain idea to
one particular sound rather than another. But ideas—at least
those that are clear and distinct—are not at all arbitrary things
depending on our fancy. (Arnauld & Nicole 1662: 28)
Plato’s Problem
Nativism provides a solution to Plato’s problem (cf.
Plato’s Meno and Theaetetus).
For it provides a science of language that shows how
an internal biological mechanism can, with little input
from the external environment (poverty of the stimuli
argument) develop (almost automatically) in each
individual the rich competence known as “knowing a
Solving Plato’s problem for language acquisition
It involves saying both what is known when one knows
a language and how one comes to know it.
We should do this with a science of the mind, not
philosophical speculations.
Chomsky vs. Plato
Plato appeals to myth, invoking the pre-existence of the
boy’s soul with other souls in the world of Forms
(ideas) and in going trough a process of reminiscence.
Chomsky solves it in proposing a naturalistic theory of
a biological system that makes language acquisition
virtually automatic.
General Presuppositions of Cartesian
The principle of language and natural logic are known
unconsciously and they are in large part a precondition
for language acquisition rather than a matter of
institution or training.
Linguistics as a science trying to bring to light these
underlying principles becomes a branch of psychology.
Thus this art [logic or art of thinking] does not consist in finding
the mean to perform these operations, since nature alone
furnished them in giving us reason, but in reflecting on what
nature makes us do. (Arnauld & Nicole 1662: 23)
The Poverty of the Stimulus
General language-acquisition schema
linguistic data)
(Grammar consisting of
principles, parameters
and lexicon)
Language acquisition is a matter of growth and
maturation of relatively fixed principles under
appropriate external conditions and training.
Cf. growth and maturation of bones: the structure of
the bones is genetically programmed, yet it needs
exercise to develop.
One learns a language because one is programmed to
learn a language, i.e. because of one’s initial state, UG.
The process of creolization underlies what happens
when a child learns her mother tongue in normal
The same kind of linguistic genius is involved every time a child
learns his or her mother tongue. ... let us do away with the
folklore that parents teach their children language. (S. Pinker.
1994. The Language Instinct: 39)
The crux of the argument is that complex language is universal
because children actually reinvent it, generation after
generation — not because they are taught, not because they are
generally smart, not because it is useful to them, but because
they just can’t help it. (Pinker 1994: 32)
The argument of innate knowledge
It rests on the actual way children acquire their mother
It is an empirical hypothesis which posits that our brain
is genetically programmed to invent a language.
Large-scale sensory deficit seems to have limited effect on
language acquisition. Blind children acquire language as the
sighted do, even color terms and words for visual experience like
“see” and “look.” There are people who have achieved close
to normal linguistic competence with no sensory input
beyond that can be gained by placing one’s hand on
another person’s face and throat. The analytic mechanism
of the language faculty seem to be triggered in much the
same way whether the input is auditory, visual, even
tactual, and seem to be localized in the same brain areas,
somewhat surprisingly.
These examples of impoverished input indicate the richness
of innate endowment — though normal language acquisition
is remarkable enough, as even lexical access shows, not only
because of its rapidity and the intricacy of result. Thus very
young children can determine the meaning of a nonsense word
from syntactic information in a sentence far more complex that
they can produce.
A plausible assumption today is that the principles of
language are fixed and innate, and that variations is restricted
in the manner indicated. Each language, then, is (virtually)
determined by a choice of values for lexical parameters: with the
array of choices, we should be able to deduce Hungarian; with
another, Yoruba. … The conditions of language acquisition
make it plain that the process must be largely inner-directed,
as in other aspects of growth, which means that all
languages must be close to identical, largely fixed by initial
state. (Chomsky 2000. New Horizons … : 121-2)
The paradox of language acquisition
[A]n entire community of highly trained professionals, bringing
to bear years of conscious attention and sharing of information,
has been unable to duplicate the feat that every normal child
accomplishes by the age of ten or so, unconsciously and
unaided. (Jackendoff 1994: 26)
Language Perception and Understanding
Perception of speech rests on innate discriminatory
There is a fundamental difference between the
perception of speech and the perception of
unarticulated sounds.
Speech perception, unlike visual perception for
instance, requires the activation of the generative rules
playing the role in the production of speech.
Both the perceptual mechanism and the mechanism of
speech production make use of the same underlying
system of generative rules.
It is because these underlying systems are the same
among us that communication can occur.
It is because of this uniformity of human nature that
we talk the way we do and succeed in understanding
each others (cf. Humboldt 1836).
Every young child (raised in an English speaking
community) would know that in English ‘blug’ is
phonetically possible while ‘bkr’ is not. And they know
it without being told.
Science of Intelligent Behaviour
It may be within the boundary of some other cognitive
beings (Martians, God) but it transcends human
Reasons vs. causes
Wittgenstein (Blue Book) says that in explaining action in
terms of their coherence and appropriateness with
respect to human aims etc. we “give reasons”, not “give
When talking about creative linguistic actions Chomsky
and Descartes seem to accept Wittgenstein’s view in
assuming that we are giving reasons, not causes.
Descartes’ dualism
It was a scientific hypothesis dictated by the science of
his time (mechanism).
Descartes did not have at his disposal the biological
science of our time, he did not know of genetic
transmission and could not possibly imagine how
human cognition can rest to such an extent on a
biological base of concept and structure acquisition.
Descartes could not imagine that these biological
mechanisms need only a little input to produce rich
conceptual material.
At present little is known on how UG is embodied in
the brain.
UG is considered as a computational system in the
head, but we do not know about the specific operations
of the brain itself and what leads to the development
of these computational systems.
A plausible view is that language is a distinct and
specific part of the human mind and not a
manifestation of a more general capacity or ability (of
general intelligence).
Linguistic capacity rests on a specific module.
It is not the sub-product of a general cognitive capacity.
People can “lose their intelligence” and yet they do not
loose their language: substantial retarded children (e.g.
Williams syndrome) manifest a good grammatical and
linguistic competence.
On the other hand, highly intelligent people may lack
linguistic capacity (e.g. aphasia).
The fact that two kinds of abilities can dissociate quantitatively
and along multiple dimensions shows that they are not
manifestations of a single underlying ability. (Pinker 2003: 23)

7 Universal Grammar - Carleton University