The Making of the Mind: The Neuroscience of Human Nature (12 page)

A final example is that some words refer to concepts that do not have a well-defined extension that can be enumerated. Take, for example, the concept
of liberty. The concept itself seems to exist chiefly in the minds of human beings. Its referential extension is as abstract as the concept itself. One can imagine counting up all the dogs in the world today—an onerous task to be sure, which would yield a very large number, but at least it is a finite number. By contrast, it is difficult to conceive of enumerating all possible examples of liberty. A richly diverse and seemingly infinite set of exemplars all fit the word well. Or consider imaginary objects. Lilliputians, hobbits, unicorns, fairies, and the like are all perfectly useful concepts with an indefinitely large extension. But they can be seen only with the mind's eye. They cannot be tracked down and counted in the physical world.

Thus, language uses abstract symbols to convey meaning. Humans use words, or patterns of sound, to refer to objects, events, beliefs, desires, feelings, and intentions. The words carry meanings. If your friend says he is happy, then you interpret this to mean something about his emotional state. If, instead of speaking, your friend whistles a tune, then his behavior may say something about his emotional state, but it is less meaningful. Your friend might whistle by habit or whistle when he is angry, sad, or happy. Unlike speech, whistling is not specialized to convey a clear meaning. Once humans learn a word, they can retrieve its mental representation, hold it in working memory, and use it in thought. The word itself is represented separately from the object or event to which it refers.

At the core of language, then, is an ability to use symbolic representation both in communicating with other human beings and in thought. As Terrence Deacon put it in
The Symbolic Species
, “language is not merely a mode of communication, it is also the outward expression of an unusual mode of thought—symbolic representation…symbolic thought does not come innately built in, but develops by internalizing the symbolic process that underlies language.”
¹²

The symbols of language are invented by human beings in such a form that they can be passed down to the next generation. Language thus must be learnable by children. Although languages can be remarkably complicated, they cannot be so complicated that children fail to acquire them and pass them on to their children. As with other cultural artifacts, language can be modified so that is evolves over multiple generations. For example, after a bit
more than four hundred years, or twenty generations, the Elizabethan English of Shakespeare's plays is recognizable but distinct from the British English spoken today. Modern readers of Chaucer's poetry in Middle English, from the fourteenth century, must struggle still more to recognize their language, while the Old English poem
Beowulf
, from the early eighth century, strikes the eye and ear as a foreign language.

The speed of linguistic change through historical time can be dauntingly fast. Merritt Ruhlen illustrated this point in
The Origin of Language
with the example of words used to express approval in American English with “the succession of ‘neat’ (1950s), ‘cool’ (1960s), ‘bad’ (1970s), ‘rad’ (1980s), and ‘awesome’ (1990s).”
¹³
All of these were existing words in the language (“rad” being an abbreviation of “radical”), but entirely new words can also be invented and find their way into the language community. For example, our modern word
dog
came from a new word,
docga
, invented in the Old English of one thousand years ago. This novel entry into the language competed with a much more ancient term found in Proto-Indo-European, whose modern equivalent is
hound
. Over the past millennium,
dog
eventually replaced
hound
as the everyday label for this commonplace category of animals. Although the
hound
remained in the language, it evolved over the generations to a narrow meaning of “hunting dog.”
¹⁴

Language can be used to communicate factual information, but this is not its sole function, and it may not even be the motivation for its origin. Venting emotions, joke telling, and social greetings all are common uses of language that bond people together in a world of shared experience, yet they are not aimed at the articulation of facts or even opinions. Moreover, some factual knowledge cannot easily be rendered in language. For example, try to describe to a friend how to tie a necktie without resorting to gestures or drawing a diagram. Or try instructing someone, in words alone, how to navigate by walking or by car to a location in a large city a mile away. Now how much easier is it if you can point to a map? These are examples of visual-spatial knowledge that is not readily encoded in words alone.

Emotions, too, can be too dense to find clear expression in strings of words. All human beings can quickly communicate joy, sadness, fear, or surprise through inborn species-specific facial expressions. These facial gestures
are universally understood across all human cultures. Yet sharing these emotions clearly through language challenges the average speaker; it is more a task for the advanced skills of poets. Consider romantic love. How does one say “I love you” in a way that it is understood through language alone, without all the other behavioral ways that human beings express their love for one another? Countless love letters, sonnets, and songs show that even the most skilled with language struggle with the expression of love.

By combining the symbols of language, any human being can say any number of things—the scope of possible expressions is inexhaustible. Consider, for example, a six-word sentence. Suppose that one were to select one of ten possible words for the first word of the sentence, one of another set of ten possible words for the second word, and so on. The number of unique sentences that could be generated following this procedure would equal 10
⁶
or one million sentences. Because you are not limited to only six-word sentences or to ten possible choices, the number of unique sentences that you might utter is infinite. Thus, abstractions of language are profoundly productive—through them the human mind can generate novel sentences endlessly.

THE MACHINERY OF LANGUAGE

How then is the productive capacity of language realized? The first part of the machinery of language is semantics, or the meanings of all its symbols. A whole word is constituted from brief fragments of sound that distinguish one meaningful word from another. For example,
bill
and
kill
convey different meanings because they differ in their initial sound. Similarly,
braised
and
praised
also differ, but here the difference in pronouncing
b
versus
p
is a very subtle one. The sounds are so close that languages related in their historical evolution often adopt only one or the other to mark meaning shifts. To illustrate this comparison, “language A might have
aba
for ‘fish,’
uda
‘walk,’ and
paga
‘tree,’ while the closely related languages B, C, and D all have
apa, uta, and paka
for the same three words.”
¹⁵

These sounds are the building blocks of meaningful components, the morphemes. A
morpheme
is a minimal unit of speech used repeatedly in a language to code a specific meaning.
¹⁶
A word such as
bill
is a morpheme, but
so, too, are prefixes and suffixes, such as
pre-
and
-es.
Each morpheme signals a distinct meaning. The suffix
-ed
on the end of a verb tells us that the action took place in the past. So, the word
billed
is composed of two morphemes, each of which conveys a specific meaning.

All the morphemes in a language, taken together, make up a mental lexicon, or the dictionary of long-term memory that humans rely on in speaking and listening and in reading and writing. Each morpheme is a lexical entry in this dictionary of the mind. In particular, one is concerned in semantics with content words, that is, the verbs and the nouns. Function words, such as articles (e.g.,
the
) and prepositions (e.g.,
by
), often serve more of a grammatical function rather than a semantic role. For example, the statement “The telephone company bills by the month” would mean the same thing if the prepositional phrase
by the month
were replaced by the adverb
monthly
. However, if
month
were replaced by
week
, then the meaning of the sentence would change to a more frequent billing cycle. Replacing the content word
telephone
with
electric
changes meaning, too, altering where one should send the payment.

The sound of a language is influenced by its phonology, the sound segments that make up the words. Some phonological segments make a difference in meaning in a given language, and these are called
phonemes
. English makes use of forty-six phonemes, some of which are used in other languages and some of which are not.
¹⁷
These include the consonants, the vowels, and combinations of vowel sounds known as
diphthongs
. For example, the consonants
l
and
r
signal different morphemes or word meanings, such as the difference between
look
and
rook
or between
lip
and
rip
. However, in Japanese this phonemic distinction is not made and the two sounds can be easily confused when native Japanese speakers learn English. Similarly, there are sounds in other languages that are not employed in English. In Spanish, the rolled
r
is articulated near the front of the mouth and is slightly different from the rolled
r
of French, which is articulated further back. Neither of these methods of articulation is used in pronouncing an English
r.

Languages also differ in the sequence of phonemes that are permitted. In English, one never sees the sequence
pt at
the beginning of a word, whereas in Greek the combination is common, as can be seen in borrowed words, such
as
pteropod
or
pterosaur
. The rules of phonology are learned implicitly through repeated exposure to a particular language. Although we are not consciously aware of these rules, we can easily decide whether a nonsense word is possible (
patik
) or impossible (
ptkia
) by making use of what we have learned unconsciously about English phonology.

The sound structure of a word is detached from its semantics in terms of the location in the brain where each is processed. In the inferior region of the left prefrontal cortex, one module specializes in processing the meaning of words whereas another is dedicated to their sounds.
¹⁸
If you think about a word that is semantically related to another word (e.g.,
chair
is related to
table
), or think about whether a word is abstract in meaning (e.g., f
reedom
) versus something concrete and easy to visualize in the mind's eye (e.g.,
tree
), then a semantic network in the inferior left prefrontal cortex is activated. If, on the other hand, you think about how many syllables a word has or whether it rhymes with another word, then a distinct phonological network is activated. This phonological network is the same as that found in the phonological loop of verbal working memory. It is located somewhat posterior in the left prefrontal cortex from the semantic network.

Sometimes during speech production, the sound of the word we want to say cannot be retrieved. There is a feeling of knowing what needs to be said, but sound structure is temporarily lost. Bits and pieces might be accessible—such as knowing the first letter or the number of syllables—but the full phonology cannot be retrieved. This frustrating experience is called a
tip-of-the-tongue state
.
¹⁹
It illustrates that the semantic representation of a word—its meaning—is stored separately from its phonological representation—how it sounds when pronounced. Unless both meaning and sound structure are jointly retrieved, the word cannot be named.

Spoken language unfolds in a sequence over time. One word follows another in temporal order. Whereas a visual event can be simultaneously grasped as a whole, the sound structure of language or any other auditory source of information, such as music, has a temporal structure. Syntax refers to the grammatical rules that specify how words and other morphemes are arranged in a sequence to yield acceptable sentences.

For example, consider this arrangement of words: “Young children learn
language very rapidly.” This string of words follows the syntactic rules of English; the sentence begins with an animate agent (“children”) that serves as its subject, followed by a verb that describes what the subject does (“learn”), which in turn is followed by the direct object of the verb (“language”). The subject is modified appropriately by placing an adjective (“young”) before the subject; reversing this word order creates a syntactic error that sounds peculiar to the ear of the listener (“Children young…”). However, there is more flexibility regarding where to position the adverbs; thus, “Young children very rapidly learn language” is also grammatical. The morpheme
-s
must be added to the verb if the subject is changed to singular rather than plural (“A young child learns…”) for the sentence to be grammatical.

When children learn their native language, they acquire implicitly an understanding of the grammatical rules and exceptions to those rules. Implicit learning occurs without conscious awareness of the rules. Just as the allowable sequences of phonemes are learned and used unconsciously, so are the permissible orders of morphemes. This refers not just to the order of whole words, but also to the use of morphemes as word endings, such as adding -
ed
to a verb to express that the action took place in the past. Knowing the syntax of a language implies the ability to decide whether a string of words is grammatical and whether two different sentences mean the same thing. For example, “Language is learned by children very rapidly” is easily recognized as semantically equivalent to our earlier sentence despite the difference in syntax.