Language in a New Key

Abstract

Much of Noam Chomsky's revolution in linguistics—including its account of the way we learn languages—is being overturned

Full text

70 Scientific American, November 2016
LANGUAGE
LINGUISTICS

November 2016, ScientificAmerican.com 71
KEY
T
             —
famously espoused by Noam Chomsky of the Massachusetts Institute of Technology—
has dominated linguistics for almost half a century. Recently, though, cognitive scien-
tists and linguists have abandoned Chomsky’s “universal grammar” theory in droves
because of new research examining many dierent languages—and the way young chil-
dren learn to understand and speak the tongues of their communities. That work fails
to support Chomsky’s assertions.
The research suggests a radically dierent view, in which
learning of a child’s first language does not rely on an innate gram-
mar module. Instead the new research shows that young children
use various types of thinking that may not be specific to language
at all—such as the ability to classify the world into categories
(people or objects, for instance) and to understand the relations
Illustration by Owen Gildersleeve
Much of
Noam Chomsky’s
revolution in
linguistics—
including
its account
of the way we
learn languages—
is being overturned
By Paul Ibbotson
and Michael Tomasello
LANGUAGE
LINGUISTICS
IN A NE W
IN BRIEF
Noam Chomsky has been a towering
giant in the eld of linguistics for many
decades, famed for his well-known the-
ory of universal grammar.
Chomsky’s idea of a brain wired with
a mental template for grammar has been
questioned, based on a lack of evidence
from eld studies of languages.
The theory has changed several times
to account for exceptions that run coun-
ter to its original postulations—marking
a retreat from its ambitious origins.
Alternatives to universal grammar
posit that children learning language
use general cognitive abilities and the
reading of other people’s intentions.

72 Scientific American, November 2016
Paul Ibbotson is a lecturer in language development
at the Open University, based in England.
Michael Tomasello is co-director of the Max Planck
Institute for Evolutionary Anthropology in Leipzig, Germany,
and author, most recently, of A Natural History of Human
Morality (Harvard University Press, 2016).
among things. These capabilities, coupled with a unique hu man
ability to grasp what others intend to communicate, allow lan-
guage to happen. The new findings indicate that if researchers
truly want to understand how children, and others, learn languag-
es, they need to look outside of Chomsky’s theory for guidance.
This conclusion is important because the study of language
plays a central role in diverse disciplines—from poetry to artifi-
cial intelligence to linguistics itself; misguided methods lead to
questionable results. Further, language is used by humans in
ways no animal can match; if you understand what language is,
you comprehend a little bit more about human nature.
Chomsky’s first version of his theory, put forward in the mid-
20th century, meshed with two emerging trends in Western intel-
lectual life. First, he posited that the languages people use to com-
municate in everyday life behaved like mathematically based lan-
guages of the newly emerging field of computer science. His
research looked for the underlying computational structure of
language and proposed a set of procedures that would create
“well-formed” sentences. The revolutionary idea was that a com-
puterlike program could produce sentences real people thought
were grammatical. That program could also purportedly explain
the way people generated their sentences. This way of talking
about language resonated with many scholars eager to em brace a
computational approach to. .. well .. . everything.
As Chomsky was developing his computational theories, he
was simultaneously proposing that they were rooted in human
biology. In the second half of the 20th century, it was becoming
ever clearer that our unique evolutionary history was responsi-
ble for many aspects of our unique human psychology, and so
the theory resonated on that level as well. His universal gram-
mar was put forward as an innate component of the human
mind—and it promised to reveal the deep biological underpin-
nings of the world’s 6,000-plus human languages. The most
powerful, not to mention the most beautiful, theories in science
reveal hidden unity underneath surface diversity, and so this
theory held immediate appeal.
But evidence has overtaken Chomsky’s theory, which has
been inching toward a slow death for years. It is dying so slowly
because, as physicist Max Planck once noted, older scholars
tend to hang on to the old ways: “Science progresses one funer-
al at a time.”
IN THE BEGINNING
   of universal grammar in the 1960s
took the underlying structure of “standard average European”
languages as their starting point—the ones spoken by most of
the linguists working on them. Thus, the universal grammar
program operated on chunks of language, such as noun phrases
(“The nice dogs”) and verb phrases (“like cats”).
Fairly soon, however, linguistic comparisons among multiple
languages began rolling in that did not fit with this neat schema.
Some native Australian languages, such as Warlpiri, had gram-
matical elements scattered all over the sentence—noun and verb
phrases that were not “neatly packaged” so that they could be
plugged into Chomsky’s universal grammar—and some sentenc-
es had no verb phrase at all.
These so-called outliers were dicult to reconcile with the
universal grammar that was built on examples from European
languages. Other exceptions to Chomsky’s theory came from the
study of “ergative” languages, such as Basque or Urdu, in which
the way a sentence subject is used is very dierent from that in
many European languages, again challenging the idea of a uni-
versal grammar.
These findings, along with theoretical linguistic work, led
Chomsky and his followers to a wholesale revision of the notion of
universal grammar during the 1980s. The new version of the theo-
ry, called principles and parameters, replaced a single universal
grammar for all the world’s languages with a set of “universal”
principles governing the structure of language. These principles
manifested themselves dierently in each language. An analogy
might be that we are all born with a basic set of tastes (sweet, sour,
bitter, salty and umami) that interact with culture, history and
geography to produce the present-day variations in world cuisine.
The principles and parameters were a linguistic analogy to tastes.
They interacted with culture (whether a child was learning Japa-
nese or English) to produce today’s variation in languages as well
as defined the set of human languages that were possible.
Languages such as Spanish form fully grammatical sentenc-
es without the need for separate subjects—for example, Tengo
zapatos (“I have shoes”), in which the person who has the shoes,
“I,” is indicated not by a separate word but by the “o” at the end
of the verb. Chomsky contended that as soon as children
encountered a few sentences of this type, their brains would set
a switch to “on,” indicating that the sentence subject should be
dropped. Then they would know that they could drop the sub-
ject in all their sentences.
The “subject-drop” parameter supposedly also determined
other structural features of the language. This notion of universal
principles fits many European languages reasonably well. But
data from non-European languages turned out not to fit the
revised version of Chomsky’s theory. Indeed, the research that
had at tempted to identify parameters, such as the subject-drop,
ultimately led to the abandonment of the second incarnation of
universal grammar because of its failure to stand up to scrutiny.
More recently, in a famous paper published in Science in
2002, Chomsky and his co-authors described a universal gram-
mar that included only one feature, called computational recur-
sion (although many advocates of universal grammar still prefer
to assume there are many universal principles and parameters).
This new shift permitted a limited number of words and rules to
be combined to make an unlimited number of sentences.
The endless possibilities exist because of the way recursion
embeds a phrase within another phrase of the same type. For
example, English can embed phrases to the right (“John hopes
Mary knows Peter is lying”) or embed centrally (“The dog that the
cat that the boy saw chased barked”). In theory, it is possible to go

November 2016, ScientificAmerican.com 73Illustration by Lucy Reading-Ikkanda
on embedding these phases infinitely. In practice, understanding
starts to break down when the phrases are stacked on top of one
another as in these examples. Chomsky thought this breakdown
was not directly related to language per se. Rather it was a limita-
tion of human memory. More important, Chomsky proposed that
this recursive ability is what sets language apart from other types
of thinking such as categorization and perceiving the relations
among things. He also proposed recently this ability arose from a
single genetic mutation that occurred be tween 100,000 and
50,000 years ago.
As before, when linguists actually went looking at the varia-
tion in languages across the world, they found counterexamples
to the claim that this type of recursion was an essential property
of language. Some languages—the Amazonian Pirahã, for in -
stance—seem to get by without Chomskyan recursion.
As with all linguistic theories, Chomsky’s universal grammar
tries to perform a balancing act. The theory has to be simple
enough to be worth having. That is, it must predict some things
that are not in the theory itself (otherwise it is just a
list of facts). But neither can the theory be so sim-
ple that it cannot explain things it should. Take
Chomsky’s idea that sentences in all the world’s
languages have a “subject.” The problem is the
concept of a subject is more like a “family
resemblance” of features than a neat category.
About 30 dierent grammatical features define
the characteristics of a subject. Any one language
will have only a subset of these features—and the sub-
sets often do not overlap with those of other languages.
Chomsky tried to define the components of the essential tool
kit of language—the kinds of mental machinery that allow hu -
man language to happen. Where counterexamples have been
found, some Chomsky defenders have responded that just be -
cause a language lacks a certain tool—recursion, for example—
does not mean that it is not in the tool kit. In the same way, just
because a culture lacks salt to season food does not mean salty is
not in its basic taste repertoire. Unfortunately, this line of reason-
ing makes Chomsky’s proposals dicult to test in practice, and in
places they verge on the unfalsifiable.
DEATH KNELLS
   in Chomsky’s theories is that when applied to language
learning, they stipulate that young children come equipped with
the capacity to form sentences using abstract grammatical rules.
(The precise ones depend on which version of the theory is in -
voked.) Yet much research now shows that language acquisition
does not take place this way. Rather young children begin by
learning simple grammatical patterns; then, gradually, they intu-
it the rules behind them bit by bit.
Thus, young children initially speak with only concrete and
simple grammatical constructions based on specific patterns of
words: “Where’s the X?”; “I wanna X”; “More X”; “It’s an X”; “I’m
X-ing it”; “Put X here”; “Mommy’s X-ing it”; “Let’s X it”; “Throw
X”; “X gone”; “Mommy X”; “I Xed it”; “Sit on the X”; “Open X”;
“X here”; “There’s an X”; “X broken.” Later, children combine
these early patterns into more complex ones, such as “Where’s
the X that Mommy Xed?”
Many proponents of universal grammar accept this charac-
terization of children’s early grammatical development. But then
THEORIES OF LANGUAGE
Noam-enclature
Noam Chomsky took the linguistics community by storm more
than 50 years ago. The idea was simple. Underlying language is
a set of rules innate to every child that generates grammatical
sentences from the earliest age. Chomsky set out to dene those
rules and how they work. Without this universal grammar, he
thought, it would be impossible for a child to learn any language.
In the ensuing years, Chomsky’s theory has gradually been chal-
lenged by new theories asserting that language is acquired as
children discern patterns in the language they hear around them.
Sentence
Noun
phrase
Verb
phrase
Determiner
Adjective
Noun
Verb
Noun
Noun
phrase
the nice dogs like
cats
wants
?
“wants”
“The dog
wants the ball.”
“The dog
wants food.”
Chomsky’s Universal Grammar
Chomsky’s universal grammar equipped the child with rules that worked on
phrases (“the nice dogs”) and rules for transforming those phrases (“Cats
are liked by the nice dogs”). The theory has evolved in recent years but still
retains the essential idea that children are born with the ability to make words
conform to a grammatical template.
Usage-Based Learning
New approaches to linguistics and psychology suggest that children’s natural
ability to intuit what others think, combined with powerful learning mechanisms
in the developing brain, diminishes the need for a universal grammar. Through
listening, the child learns patterns of usage that can be applied to dierent
sentences. The word “food” might replace the word “ball” after the phrase
“The dog wants.” Studies show that this theory of building up knowledge
of word meaning and grammar approximates the way that two- and three-
year-olds actually learn language.
Sentence
Noun
phrase
Verb
phrase
Determiner
Adjective
Noun
Verb
Noun
Noun
phrase
the nice dogs like
cats
wants
?
“wants”
“The dog
wants the ball.”
“The dog
wants food.”
The
brain’s innate
sentence-diagramming
machine, according to
Chomsky, would t words
into correct grammatical
slots—“nice” (adjective),
“dogs” (noun).

they assume that when more complex constructions emerge,
this new stage reflects the maturing of a cognitive capacity that
uses universal grammar and its abstract grammatical categories
and principles.
For example, most universal grammar approaches postulate
that a child forms a question by following a set of rules based on
grammatical categories such as “What (object) did (auxiliary)
you (subject) lose (verb)?” Answer: “I (subject) lost (verb) some-
thing (object).” If this postulate is correct, then at a given devel-
opmental period children should make similar errors across all
wh-question sentences alike. But children’s errors do not fit this
prediction. Many of them early in development make errors
such as “Why he can’t come?” but at the same time as they make
this error—failing to put the “can’t” before the “he”—they cor-
rectly form other questions with other “wh-words” and auxilia-
ry verbs, such as the sentence “What does he want?”
Experimental studies confirm that children produce correct
question sentences most often with particular wh-words and aux-
iliary verbs (often those with which they have most experience,
such as “What does .. . ”), while continuing to make errors with
question sentences containing other (often less frequent) combi-
nations of wh-words and auxiliary verbs: “Why he can’t come?”
The main response of universal grammarians to such find-
ings is that children have the competence with grammar but that
other factors can impede their performance and thus both hide
the true nature of their grammar and get in the way of studying
the “pure” grammar posited by Chomsky’s linguistics. Among
the factors that mask the underlying grammar, they say, include
immature memory, attention and social capacities.
Yet the Chomskyan interpretation of the children’s behavior is
not the only possibility. Memory, attention and social abilities
may not mask the true status of grammar; rather they may well
be integral to building a language in the first place. For example,
a recent study co-authored by one of us (Ibbotson) showed that
children’s ability to produce a correct irregular past tense verb—
such as “Every day I fly, yesterday I flew” (not “flyed”)—was asso-
ciated with their ability to inhibit a tempting response that was
unrelated to grammar. (For example, to say the word “moon”
while looking at a picture of the sun.) Rather than memory, men-
tal analogies, attention and reasoning about social situations get-
ting in the way of children expressing the pure grammar of
Chomskyan linguistics, those mental faculties may explain why
language develops as it does.
As with the retreat from the cross-linguistic data and the
tool-kit argument, the idea of performance masking compe-
tence is also pretty much unfalsifiable. Retreats to this type of
claim are common in declining scientific paradigms that lack a
strong em pirical base—consider, for instance, Freudian psy-
chology and Marxist in terpretations of history.
Even beyond these empirical challenges to universal grammar,
psycholinguists who work with children have diculty conceiving
theoretically of a process in which children start with the same
algebraic grammatical rules for all languages and then proceed to
figure out how a particular language—whether English or Swahi-
li—connects with that rule scheme. Linguists call this conundrum
the linking problem, and a rare systematic attempt to solve it in
the context of universal grammar was made by Harvard Universi-
ty psychologist Steven Pinker for sentence subjects. Pinker’s ac -
count, however, turned out not to agree with data from child de -
velopment studies or to be applicable to grammatical categories
other than subjects. And so the linking problem—which should be
the central problem in applying universal grammar to language
learning—has never been solved or even seriously confronted.
AN ALTERNATIVE VIEW
    ineluctably to the view that
the notion of universal grammar is plain
wrong. Of course, scientists never give up on
their favorite theory, even in the face of con-
tradictory evidence, until a reasonable alter-
native appears. Such an alternative, called
usage-based linguistics, has now arrived. The
theory, which takes a number of forms, pro-
poses that grammatical structure is not in -
nate. Instead grammar is the product of his-
tory (the processes that shape how languages
are passed from one generation to the next) and human psychol-
ogy (the set of social and cognitive capacities that allow genera-
tions to learn a language in the first place). More important, this
theory proposes that language recruits brain systems that may
not have evolved specifically for that purpose and so is a dier-
ent idea to Chomsky’s single-gene mutation for recursion.
In the new usage-based approach (which includes ideas from
functional linguistics, cognitive linguistics and construction
grammar), children are not born with a universal, dedicated tool
for learning grammar. Instead they inherit the mental equiva-
lent of a Swiss Army knife: a set of general-purpose tools—such
as categorization, the reading of communicative intentions, and
analogy making, with which children build grammatical catego-
ries and rules from the language they hear around them.
For instance, English-speaking children understand “The cat
ate the rabbit,” and by analogy they also understand “The goat
tickled the fairy.” They generalize from hearing one example to
another. After enough examples of this kind, they might even be
able to guess who did what to whom in the sentence “The gazzer
mibbed the toma,” even though some of the words are literally
nonsensical. The grammar must be something they discern
beyond the words themselves, given that the sentences share lit-
tle in common at the word level.
The meaning in language emerges through an interaction
between the potential meaning of the words themselves (such
as the things that the word “ate” can mean) and the meaning of
the grammatical construction into which they are plugged. For
example, even though “sneeze” is in the dictionary as an intran-
sitive verb that only goes with a single actor (the one who sneez-
es), if one forces it into a ditransitive construction—one able to
In the new usage-based approach,
children are not born with a universal,
dedicated tool for the learning of
grammar. Instead they inherit the
mental equivalent of a Swiss Army knife.
Watch Tomasello give a talk on human communication at ScienticAmerican.com/nov2016/tomasello
SCIENTIFIC AMERICAN ONLINE

November 2016, ScientificAmerican.com 75
take both a direct and indirect object—the result might be “She
sneezed him the napkin,” in which “sneeze” is construed as an
action of transfer (that is to say, she made the napkin go to him).
The sentence shows that grammatical structure can make as
strong a contribution to the meaning of the utterance as do the
words. Contrast this idea with that of Chomsky, who argued
there are levels of grammar that are free of meaning entirely.
The concept of the Swiss Army knife also explains language
learning without any need to invoke two phenomena required by
the universal grammar theory. One is a series of algebraic rules for
combining symbols—a so-called core grammar hardwired in the
brain. The second is a lexicon—a list of exceptions that cover all of
the other idioms and idiosyncrasies of natural languages that
must be learned. The problem with this dual-route approach is
that some grammatical constructions are partially rule-based and
also partially not—for example, “Him a presidential candidate?!”
in which the subject “him” retains the form of a direct object but
with the elements of the sentence not in the proper order. A native
English speaker can generate an infinite variety of sentences using
the same approach: “Her go to ballet?!” or “That guy a doctor?!” So
the question becomes, are these utterances part of the core gram-
mar or the list of exceptions? If they are not part of a core gram-
mar, then they must be learned individually as separate items.
But if children can learn these part-rule, part-exception utteranc-
es, then why can they not learn the rest of language the same
way? In other words, why do they need universal grammar at all?
In fact, the idea of universal grammar contradicts evidence
showing that children learn language through social interaction
and gain practice using sentence constructions that have been
created by linguistic communities over time. In some cases, we
have good data on exactly how such learning happens. For exam-
ple, relative clauses are quite common in the world’s languages
and often derive from a meshing of separate sentences. Thus, we
might say, “My brother. . . . He lives over in Arkansas . . . . He likes
to play piano.” Because of various cognitive-processing mecha-
nisms—with names such as schematization, habituation, decon-
textualization and automatization—these phrases evolve over
long periods into a more complex construction: “My brother,
who lives over in Arkansas, likes to play the piano.” Or they
might turn sentences such as “I pulled the door, and it shut”
gradually into “I pulled the door shut.”
What is more, we seem to have a species-specific ability to de -
code others’ communicative intentions—what a speaker intends to
say. For example, I could say, “She gave/bequeathed/sent/loaned/
sold the library some books” but not “She donated the library
some books.” Recent research has shown that there are several
mechanisms that lead children to constrain these types of inap-
propriate analogies. For example, children do not make analogies
that make no sense. So they would never be tempted to say “She
ate the library some books.” In addition, if children hear quite
often “She donated some books to the library,” then this usage pre-
empts the temptation to say “She donated the library some books.”
Such constraining mechanisms vastly cut down the possible
analogies a child could make to those that align the communica-
tive intentions of the person he or she is trying to understand.
We all use this kind of intention reading when we understand
“Can you open the door for me?” as a request for help rather
than an inquiry into door-opening abilities.
Chomsky allowed for this kind of “pragmatics”—how we use
language in context—in his general theory of how language
worked. Given how ambiguous language is, he had to. But he
appeared to treat the role of pragmatics as peripheral to the
main job of grammar. In a way, the contributions from usage-
based approaches have shifted the debate in the other direction
to how much pragmatics can do for language before speakers
need to turn to the rules of syntax.
Usage-based theories are far from oering a complete ac -
count of how language works. Meaningful generalizations that
children make from hearing spoken sentences and phrases are
not the whole story of how children construct sentences either—
there are generalizations that make sense but are not grammati-
cal (for example, “He disappeared the rabbit”). Out of all the pos-
sible meaningful yet ungrammatical generalizations children
could make, they appear to make very few. The reason seems to
be they are sensitive to the fact that the language community to
which they belong conforms to a norm and communicates an
idea in just “this way.” They strike a delicate balance, though, as
the language of children is both creative (“I goed to the shops”)
and conformative to grammatical norms (“I went to the shops”).
There is much work to be done by usage-based theorists to
explain how these forces interact in childhood in a way that
exactly explains the path of language development.
A LOOK AHEAD
   the Chomskyan paradigm was proposed, it was a rad-
ical break from the more informal approaches prevalent at the
time, and it drew attention to all the cognitive complexities in -
volved in becoming competent at speaking and understanding
language. But at the same time that theories such as Chomsky’s
allowed us to see new things, they also blinded us to other aspects
of language. In linguistics and allied fields, many researchers are
be coming ever more dissatisfied with a totally formal language
approach such as universal grammar—not to mention the empir-
ical inadequacies of the theory. Moreover, many modern re -
searchers are also unhappy with armchair theoretical analyses,
when there are large corpora of linguistic data—many now avail-
able online—that can be analyzed to test a theory.
The paradigm shift is certainly not complete, but to many it
seems that a breath of fresh air has entered the field of linguistics.
There are exciting new discoveries to be made by investigating the
details of the world’s dierent languages, how they are similar to
and dierent from one another, how they change historically, and
how young children acquire competence in one or more of them.
Universal grammar appears to have reached a final impasse.
In its place, research on usage-based linguistics can provide a
path forward for empirical studies of learning, use and histori-
cal development of the world’s 6,000 languages.
MORE TO EXPLORE
Constructing a Language: A Usage-Based Theory of Language Acquisition.
Michael Tomasello. Harvard University Press, 2003.
Constructions at Work: The Nature of Generalization in Language. Adele Gold-
berg. Oxford University Press, 2006.
Language, Usage and Cognition. Joan Bybee. Cambridge University Press, 2010.
FROM OUR ARCHIVES
The “It” Factor. Gary Stix; September 2014.
scientificamerican.com/magazine/sa