INTRODUCTION TO LINGUISTICS (2019)

LECTURE TWO (12 pages)

THE STUDY OF LANGUAGE
SUMMARY
Languages began to be studied a very long time ago (about in the 5th century BC), though it
was in the 19th century that historical language study began to meet the criteria of
scientificness. Earlier language study can be called Traditional Grammar.
Traditional Grammar was not sufficiently scientific.
1) It was not explicit enough: it was often too vague in its statements and its definitions
were often too loose. For example, the noun was defined as ‘the name of a person,
place or thing’. (e.g. reflection belongs to none of them)
2) It was not systematic enough: it ignored spoken language and was preoccupied with
written language, especially with the written language of older literary works.
3) It was not objective enough: it was often prescriptive rather than descriptive, i.e.
instead of recording what the language examined was like, traditional grammarians
often tried to prescribe what it should be like. In these attempts they relied on analogy
with Latin. (e.g. they argued that split infinitive to humbly apologize was wrong and it
should be to apologize humbly).
Comparative Philology was the dominant kind of language study in the 19th century. It
emerged after the discovery that Sanskrit was related to Latin and Greek. This discovery was
made in 1786 by Sir William Jones, a British government official working in India.
Comparative Philology narrowed down the concept of language study to a study of the
history and genetical relationships of languages and of the written records that were available.
Nevertheless, throughout the 19th century various language families and branches were
discovered, a Proto-Indo-European parent language was reconstructed. In the last quarter of
the 19th century, a group of scholars in and around Leipzig, nicknamed the
Neogrammarians, claimed that language changes were not just accidental events or optional
tendencies, but ‘laws’.
Modern Linguistics emerged almost simultaneously in Europe and the USA in the early
decades of the 20th century. That time the study of language was characterized by two
features: the inheritance of a long period of Traditional Grammar, and the predominantly
historical interest of 19th century Comparative Philology. Modern linguistics appeared as a
kind of revolt against this background.
Ferdinand de Saussure, a Swiss scholar (a professor of Sanskrit at the University of
Geneva) was the first great figure of modern linguistics in Europe, whose ideas about
language and language study went far beyond the limitations of Comparative Philology. He
was the first to emphasize the difference between (a) language as an abstract system, residing
in the collective consciousness of the community (which he called la langue) and (b)
language as the realization of that system (which he called la parole). He separated the
synchronic and diachronic aspects of language study and argued for the primacy of the
former. According to Saussure, linguistic signs enter into two kinds of relationship:
syntagmatic and paradigmatic. The syntagmatic relationship is a linear (horizontal, chain)
relationship, which exists between the signs that follow one another in a complex unit. The
paradigmatic relationship is a vertical (choice) relationship, which exists between a sign
present in a particular environment and all the other signs that could replace it while still
yielding a well-formed complex unit.
Sapir-Whorf Hypothesis. Linguistic research in the USA also began in the early decades of
the 20th century. This linguistics was usually called structural or descriptive linguistics.
American descriptivists (descriptive linguists) tried to describe each language in its own
terms and they emphasized (even exaggerated) the differences between languages. One of
them, Martin Jones, said: “Languages differ from one another without limit and in
unpredictable ways.” This is the essence of linguistic relativism. Linguistic relativism is the
assumption that any natural language can be totally different from other languages. But some
linguists went even further. Sapir and especially Whorf thought that languages not only
differed from one another without limit but also that the language of the community
determined the way in which that community saw the world. This latter view is called
linguistic determinism. The combination of linguistic relativism and linguistic determinism
became known as the Sapir – Whorf hypothesis. According to the strong version of this
hypothesis, the individual is not free in his experience of the world, because the vocabulary
and grammatical categories of his native language determine the ways in which he can
interpret his experience. The conclusion was drawn that people belonging to different
cultural-linguistic groups not only spoke differently but also thought differently: i.e. each
cultural-linguistic community lived in the “prison” of its language. This conclusion, however,
cannot be accepted. It is true that different communities cut up reality in different ways, but
this is because different communities find different things important in their life. According
to the weak form of the Sapir – Whorf hypothesis, language influences thought. This seems
to be correct. Certain things are less codable (i.e. less expressible) in some languages than in
others. The codability of an aspect of reality in a particular language means having a word for
it, or at least the possibility of a simple paraphrase.
American Structuralist Linguistics. American structuralist linguist Leonard Bloomfield,
and his followers, the Bloomfieldians thought that a linguist should collect observable data,
i.e. real utterances, and analyze these data, and classify the physical features of the utterances
collected. The Bloomfieldians dealt with phonetics, phonology, morphology and syntax, but
rejected semantics, thinking that the study of meaning would only be possible when human
knowledge had become far more advanced. The only aspect of meaning that they paid
attention to was whether two forms (signs or sign combinations) had the same meaning or
different meanings. They used a strictly formal analysis. This was an analysis without
reference to meaning, and it was based on an examination of distribution and constituency.
The distribution of a language element (i.e. of a phoneme, or morpheme, or word) is the
sum of all the environments in which it occurs. If two language elements always occur in
different environments, they never enter into a paradigmatic relationship with each other,
they are in complementary distribution. The other important method of formal analysis
introduced by the Bloomfieldians, was constituent analysis. (they called it ‘immediate
constituent analysis’ or IC analysis). This means cutting syntactic units (or words) into their
constituents until we reach individual words (or morphemes). Cutting units into its
constituents is based on the test of substitution (replacement). Constituent analysis can be
visualized in essentially two ways, viz. by (1) bracketings, or by (2) tree diagrams.
 (1) [[My friend] [ran home]]
Constituent analysis was suitable for resolving certain ambiguities, by showing different
constituent structures, e.g.:
[old [men and women]] vs.[[old men] and women]
Generative Linguistics. However, there were lots of ambiguities which constituent analysis
could not resolve. For instance, The lamb is ready to eat. has two distinct meanings (is
ambiguous), but the American structuralists could give it only one analysis: [[The
lamb][is[ready[to eat]]]].
The growing dissatisfaction with the limitations of structuralist linguistics led to the
emergence of a radically new type of linguistic analysis towards the end of the 1950s. This
has become known as transformational-generative linguistics, or just generative linguistics
(= generative grammar), for short. This kind of analysis distinguishes two levels of syntactic
structures: a surface structure (which was recognized by the structuralists, too) and an
underlying abstract deep structure (which was not recognized by the structuralists).
Transformational-Generative grammar is transformational because it explains surface
structure as being derived from deep structure by a series of changes: transformations.
Transformational-Generative grammar is generative, because it can generate (i.e. produce,
define and explicitly characterize) all and only the grammatical sentences of a language. This
means that (a) by applying the rules of the grammar, we always get a syntactically well-
formed sentence, (b) this kind of grammar generates all the well-formed sentences of a
language, i.e. not only those, that have been uttered but also those that have not been uttered
but could be uttered, and are, thus, potential sentences of the language. The number of
possible grammatical sentences in any language at any one time is infinite, but the rules
which make this infinite variation possible are finite (otherwise the native speaker would not
be able to learn them).
The founder and most influential representative to this day of generative linguistics has been
the American linguist Noam Chomsky. According to Chomsky, a generative grammar is a
model for the native speaker’s intuitive knowledge of the language (i.e. his internal grammar)
and is genetically inherited. Chomsky calls the native speaker’s language-knowledge
competence (I- language). The actual use of language-knowledge in real-life situation is
called performance. Chomsky’s distinction between competence and performance reminds
us Saussure’s distinction between langue and parole. However, Saussure’s langue was static:
it was the system of linguistic signs. Chomsky’s competence is dynamic: it puts the
generation of sentences in the center of attention. Another difference is that Saussure thought
of langue as being the collective consciousness of a community. Chomsky thinks of
competence as knowledge whose basis is given to every normal human being by birth.
Chomsky’s conclusion is that linguistics is a branch of cognitive psychology. Thus,
generative linguistics has extended the status of psycholinguistics from being a mere branch
of linguistics, to being the dominant branch of modern linguistics.
 LECTURE
Language: Externalized and internalized
A language is a linguistic code, which its speakers know and use, and which manifests itself
in its speakers’ linguistic knowledge and in the actual utterances that its speakers make in
linguistic communication. Consequently, language can be regarded as existing in essentially
two modes. On the one hand it can be looked upon as a body of objective facts (strings of
sounds or letters) produced and perceived by its users in linguistic communication. On the
other hand, it can be regarded as the language users’ knowledge which makes linguistic
communication possible, an internal property of the human mind. One of the greatest figures
in modern linguistics, Noam Chomsky, has called these two modes of language Externalized
Language (E-language) and Internalized Language (I-language), respectively.
The dominant kind of language study in the first half of the 20th century, viz. Structuralist
Linguistics, concentrated on E-language. It aimed at collecting samples of E-language, i.e.
samples of the actual products of linguistic communication, as objects independent of the
mind, and then describing the regularities (patterns, structures) found in those samples. Since
then, however, the interest and emphasis of language study has shifted to I-language, i.e. to
the knowledge that native speakers of a language possess and use when they communicate
linguistically. Generative Linguistics aims at modelling the I-language of the native speaker,
i.e. his/her linguistic knowledge or internal grammar.

Components of language
A natural language (whether we look upon it as E-language or I-language) has several
components. The central ones are phonology, morphology, syntax, and semantics.
• Phonology includes the phonemes (basic sounds) and the discrete suprasegmental elements
(stress patterns, tones, intonation) in the language. The phonological component also contains
rules that regulate how phonemes can be combined in morphemes and words. For example,
the sequences /kot/and /tok/ are phonologically well-formed in English, but */kto/ or */tko/
are phonologically ill-formed.
• Another component is morphology. This includes the morphemes and the rules for
combining them to derive and inflect words in a particular language. In English, for instance,
the morpheme -ion can be added to the verb elect (which is a vocabulary item) and the result
is the noun election (which is a new vocabulary item derived from the former one). In a
similar way, the plural morpheme -s can be added to the noun election to obtain the plural
form of the same noun: elections (which is not a new vocabulary item but the inflected
variant of an already existing one). The morphological rules of English tell us that the
sequence un-friend-li-ness is a morphologically well-formed word, while *friend-li-un-ness is
not.
• Syntax is the component of language that contains the rules for putting together words in
phrases and phrases in sentences. For example, the English sentence He went to London. is
syntactically well-formed, whereas *To he London went. is syntactically ill-formed.
• Finally, languages also contain a system of meanings: this component is known as
semantics. The semantic rules specify which sentences are semantically normal and which
are semantically anomalous. For instance, This woman is the mother of three girls. is
semantically normal but! This woman is the father of three oil-wells. is anomalous.
In addition, we can also separate a special component in which all the central components
may play a role, viz. a lexicon. This is a list of the vocabulary items of a language and it
contains all idiosyncratic information about those vocabulary items (such as the unpredictable
aspects of their phonology, morphology, syntactic behaviour, and meaning). Words, once
formed and established as vocabulary items, are stored in the lexicon, from where they can be
retrieved as wholes and do not have to be put together again from their constituent
morphemes every time they are used by a speaker.
Native speakers of a language have linguistic knowledge: they know their language. They
possess I-language, they have an internal grammar. They know the elements and the rules
in the various components of their language, after all they use those elements and obey those
rules all the time and, on the basis of this knowledge, they can tell whether a string of words
in their language is grammatical or not. But most speakers are unable to explain to their
children or to their foreign friends why one string of words is grammatical in their language
and another is not. This is because their linguistic knowledge (internal grammar) is intuitive
(subconscious), and they cannot express it explicitly (i.e. clearly and definitely).

Linguistics and its branches

If we want to obtain explicit knowledge about language, we must study language
systematically and objectively, i.e. we must deal with linguistics. Linguistics seeks explicit
knowledge about language, by submitting it to systematic and objective study. A study that is
systematic, objective, and seeks explicit knowledge is scientific. Linguistics is the scientific
study of language (i.e. E-language and/or I-language).
The product of linguistics is an objective, systematic, and explicit account of (some aspect of)
language, i.e. an explicit grammar.
A linguist is a person who is professionally engaged in the scientific study of some aspect of
language (i.e. of one particular language or of several languages or of human language in
general). From this definition it follows that someone who knows a number of languages (i.e.
a polyglot) is not necessarily a linguist, and a linguist is not necessarily someone who knows
a number of languages.
Linguistics, or its product, a grammar, has branches corresponding to the central components
of language. Phonology is the study of the phonemes and their combinations in words and
morphemes, and also of the discrete suprasegmental elements in words and sentences.
Morphology is the study of word derivation and word inflection in terms of constituent
morphemes. Syntax is the study of sentence formation. Semantics is the study of the
meaning of words and sentences. Lexicology is the study of the lexicon, i.e. the phonological,
morphological, syntactic, and semantic properties of vocabulary items. Moreover, all these
can be studied from a synchronic point of view (how they constitute a particular state of
language at a particular point of time), or from a diachronic (historical) point of view (how
they change through time).
In a somewhat broader concept of linguistics, there are phonetic and pragmatic components,
too. Phonetics is closely related to phonology, it is the study of the production, physical
properties and perception of the actual sounds realizing the phonemes and of the
suprasegmental elements of speech. Pragmatics is close to semantics and the difference is
not always quite clear. We can say that while semantics examines what sentences and words
mean in themselves, pragmatics studies the ways in which they obtain different
interpretations when uttered in different situations. For instance, if I put the question Can you
play the piano? to a person I am interviewing in a room where there is no piano, my utterance
will count as a real yes-or-no question. But if I say the same utterance to a person who is
known to be a good pianist, and I point towards a piano at the same time, my utterance will
count as a request to play.
The scope of linguistics can be extended further. It can include sociolinguistics. This is an
interdisciplinary branch of study (relevant to both linguistics and sociology), studying the
different varieties of a language used by different geographical and socio-cultural subsections
of a community, or varieties used by the same group of speakers in different social situations.
Psycholinguistics, another interdisciplinary subject, deals with areas such as the mental
processes that take place when we produce and receive linguistic messages, or the processes
of native language acquisition. And finally, linguistics can be put in the service of a large
number of other fields, some more practical, some more theoretical, such as e.g. foreign-
language teaching, speech therapy, successful advertising, literary criticism, stylistics, etc.
These involve various kinds of applied linguistics. For example, when a doctor wants
to cure a patient who suffers from aphasia (i.e. who has lost – partly or completely – the
ability to use language), the doctor will have to know about the language system. In such
cases linguistics helps the doctor in his/her work.

Traditional Grammar
Languages began to be studied a very long time ago: in the 5th century BC or earlier, but it is
only since the 19th century that we can speak about linguistics. It was in the 19th century that
historical language study began to meet the criteria of scientificness and only in the 20th
century that the study of contemporary languages became scientific in today’s sense of the
word.
Earlier language study can be called Traditional Grammar. In principle, this kind of
language study dealt with the contemporary state of languages but it often mixed its
synchronic statements with diachronic ones.
Traditional Grammar was not sufficiently scientific. (a) It was not explicit enough: it was
often too vague in its statements and its definitions were often too loose. For example, the
noun was defined as “the name of a person, place or thing”, although there are lots of words
that we intuitively feel to be nouns even though they are not the names of persons, places or
things, e.g. reflection. (b) It was not systematic enough: it ignored spoken language and
was preoccupied with written language, especially with the written language of older literary
works. (c) It was not objective enough: it was often prescriptive and puristic rather than
descriptive, i.e. instead of recording what the language examined was like, traditional
grammarians often tried to prescribe what it should be like. In these attempts they relied on
their subjective wishes and speculations and on historical, logical and aesthetic arguments,
andon analogy with Latin. For example, they argued that the split infinitive, which is quite
common in English, was incorrect: “You shouldn’t say to humbly apologize, you should say:
to apologize humbly”. The idea that the split infinitive was wrong was based on Latin. It was
believed that, since a Latin infinitive was only one word, its English equivalent should also
be as near to one word as possible. Traditional grammarians thought that language change
was harmful and they fought against it.
With all its weaknesses, however, Traditional Grammar accumulated a great number of facts
about individual languages and elaborated linguistic terminology. Modern linguistics would
not have been born if there had been no Traditional Grammar to prepare the way for it.

Comparative Philology
Comparative Philology was the dominant kind of language study in the 19thcentury. It was
scientific in several respects. However, it narrowed down the concept of language study to a
study of the history and genetical relationships of languages and of the written records that
were available.
This kind of linguistics emerged after the discovery that Sanskrit was related to Latin and
Greek. The discovery was made in 1786, by a British government official working in India,
Sir William Jones. Throughout the 19thcentury, language scholars tried to establish genetical
relationships between languages. That was the time when the various language families and
branches were discovered, for example the Germanic branch (of which English is a member)
and a Proto-Indo-European parent language was reconstructed. In Comparative Philology the
study of language was beginning to develop towards an autonomous, independent branch of
study. Language began to be studied for its own sake. Besides, this kind of language study
had an objective method: it was based on textual evidence, i.e. E-language facts, found in
earlier written records of language, and it also tried to show language change in a systematic
way, as a process determined by rules. (In the last quarter of the 19th century, a group of
scholars in and around Leipzig, nicknamed the Neogrammarians, claimed that language
changes were not just accidental events or optional tendencies, but “laws”.)
Meanwhile, the study of the contemporary state of languages went on in the non-scientific (or
not sufficiently scientific) framework of Traditional Grammar.

The Beginnings of Modern Linguistics in Europe, Saussure

Modern linguistics emerged almost simultaneously in Europe and the USA in the early
decades of the 20th century.
In Europe the study of language at the beginning of the 20th century was characterized by
two features: the inheritance of a long period of Traditional Grammar, and the predominantly
historical interest of 19th century Comparative Philology. Modern linguistics appeared as a
kind of revolt against this background. The first great figure of modern linguistics in Europe,
Ferdinand de Saussure, a Swiss scholar, was a comparative philologist himself (a professor
of Sanskrit at the University of Geneva), but his ideas about language and language study
went far beyond the limitations of Comparative Philology.
• He was the first to emphasize the difference between (a) language as an abstract system,
residing in the collective consciousness of the community (which he called la langue) and (b)
language as the realization of that system (which he called la parole).
• He separated the synchronic and diachronic aspects of language study, and argued for the
primacy of the former by saying that the synchronic aspect deals with language as a
collection of simultaneous facts, existing as a state at a particular point of time, whereas the
diachronic regards language as a succession of states, so it is the states that have to described
first.
• According to Saussure, linguistic signs enter into two kinds of relationship: syntagmatic and
paradigmatic. The syntagmatic relationship is a linear (horizontal, chain) relationship,
which exists between the signs that follow one another in a complex unit. For example, the
four words in This coffee is strong. are in a syntagmatic relationship: they are placed one after
the other along the syntagmatic axis, and each of the words has a particular environment or
CONTEXT which consists of the other words on its left and right.
The paradigmatic relationship is a vertical (choice) relationship, which exists between a sign
present in a particular environment and all the other signs that could replace it while still
yielding a well-formed complex unit. For instance, coffee in the above sentence is in a
paradigmatic relationship with tea, student, girl, wall, light, whisky, cigar, etc.

The Beginnings of Modern Linguistics in America, the Sapir—Whorf Hypothesis

Linguistic research in the USA also began in the early decades of the 20th century, but with a
different motivation. Here it was found that the languages of the American Indian population
(the Amerindian languages) were threatened with extinction and so the main aim was to
describe these languages as quickly and accurately as possible. Modern American linguistics
in the first half of the 20thcentury was usually called structural(ist) or descriptive
linguistics.
The Amerindian languages did not make a traditional approach possible. They existed only in
a spoken form, they had no earlier written records, they were very different from most of the
languages studied until then, and the linguists who wanted to describe them did not speak
them, so no prescriptive and puristic statements could be made about them. Briefly: these
languages forced languages scholars to adopt a non-traditional approach to language, based
on objectivity, systematicness and explicitness.
American descriptivists tried to describe each language in its own terms and they emphasized
(even exaggerated) the differences between languages. One of them, Martin Jones, said:
“Languages differ from one another without limit and in unpredictable ways.” This is the
essence of linguistic relativism. Linguistic relativism is the assumption that any natural
language can be totally different from other natural languages.
But some linguists went even further. Sapir and especially Whorf thought that languages not
only differed from one another without limit but also that the language of a community
determined the way in which that community saw the world. This latter view is called
linguistic determinism. The combination of linguistic relativism and linguistic determinism
became known as the Sapir—Whorf hypothesis. According to the strong version of the
hypothesis the individual is not free in his experience of the world, because the vocabulary
and grammatical categories of his native language determine the ways in which he can
interpret his experience. For instance, the American linguist Boas discovered that in Eskimo
there are several different words for different kinds of snow, whereas in English there is only
one generic term: snow. Other linguists collected similar facts from other languages. (For
instance, the Navajo language has no separate words for blue and green but has two separate
words for different shades of black; the Hopi language does not distinguish present, past and
future tenses; in Kwakiutl the distinction between singular and plural number is not
obligatory, etc.) On the basis of such examples the conclusion was drawn that people
belonging to different cultural-linguistic groups not only spoke differently but also thought
differently: i.e. each cultural-linguistic community lived in the “prison” of its language. This
conclusion, however, cannot be accepted. It is true that different languages cut up reality in
different ways, but this is because different communities find different things important in
their life. The fact that the English have no separate words for different kinds of snow does
not mean that they cannot see these differences, only that they are not significant to them.
When these differences do become important, the English can paraphrase and say “falling
snow”, “hard packed snow”, “powdery snow”, etc. The main counter-argument against the
strong form of linguistic determinism is the possibility of translation. Translation is possible
for most of the time and although we cannot always translate everything with the same ease,
we are nevertheless usually able at least to paraphrase or explain what we mean in any
language.
However, the weak form of the Sapir-Whorf hypothesis, according to which language
influences thought, seems to be correct. Certain things are less codable (i.e. less expressible)
in some languages than in others. The codability of an aspect of reality in a particular
language means having a word for it, or at least the possibility of a simple paraphrase. People
tend to notice and remember the things that are codable in their language better than things
that are not codable .But differences in codability between languages are of secondary
importance: it is only the less basic, culture-specific concepts that may present codability
problems. The essential things are equally codable because they are equally relevant to all
cultures.

The Great Synthesis of American Structuralist Linguistics

Leonard Bloomfield, and his followers, the Bloomfieldians, thought that a linguist should
collect observable data, i.e. real utterances, and analyze these data, i.e. segment and classify
the physical features of the utterances collected. A body of such data (a set of observed and
collected utterances) is a corpus. Using a corpus for linguistic investigation is called the
“corpus-based” or inductive procedure. In Chomsky’s terminology this means that American
structuralism was preoccupied with discovering and describing the E-language aspect of
natural languages.
The Bloomfieldians dealt with phonetics, phonology, morphology, and syntax, but rejected
semantics, thinking that the study of meaning would only be possible when human
knowledge had become far more advanced. The only aspect of meaning that they paid
attention to was whether two forms (signs or sign combinations) had the same meaning or
different meanings. They used a strictly formal analysis. This was an analysis without
reference to meaning, and it was based on an examination of distribution and constituency.
• The distribution of a language element (i.e. of a phoneme or morpheme or word) is the
sum of all the environments in which it occurs. If two language elements always occur in
different environments, i.e. they occur in mutually exclusive environments, then there is not
even one environment in which one could replace the other. To put it differently: they never
enter into a paradigmatic relationship with each other. In this case we say that the two
language elements have totally different distributions: they are in complementary
distribution. This means that where one of them can occur, the other cannot occur, and vice
versa. For instance, the English phoneme /l/ has two variants, and they are in complementary
distribution.1 The “clear” variant [l] (which is like Hungarian /l/)occurs before vowels, e.g.
[hel] in Helen, and the “dark” variant […] (which is pronounced with a cupped tongue, i.e.
with a raising of the back part of the tongue) occurs elsewhere, i.e. before consonants, e.g.
[he…p] help and in word-final position, e.g. [he…] hell. By contrast, if the distributions of
two language elements are not entirely different, i.e. there is at least one common
environment in which one could replace the other, the two elements are not in
complementary distribution.
• The other important method of formal analysis which the Bloomfieldians introduced was
constituent analysis. (The Bloomfieldians themselves called it “immediate constituent
analysis” or “IC analysis”.) This means cutting syntactic units (or words) into their
constituents, then the constituents into their constituents, and so on until we reach the
individual words (or morphemes). Cutting a unit into its constituents is based on the test of
substitution (replacement). For instance, the sentence My friend ran home. can be divided
into two: [My friend] and [ran home]because My friend can be replaced by a simpler
constituent, e.g. Peter, as in Peter ran home; and because ran home can also be replaced by a
simpler constituent, e.g. slept, as in My friend slept. So we divide the sentence into [My
friend] and[ran home], and then, through further applications of the substitution test, these
parts can be divided into even smaller constituents. Constituent analysis can be visualized in
essentially two ways, viz. by bracketings, or by tree diagrams.
(1) [S[NP[Det My][N friend]] [VP[V ran] [Adv home]]].
The constituents in the representations in (1) and (2) are labelled, S stands for Sentence, NP
for Noun Phrase, VP for Verb Phrase, Det for Determiner, N for Noun, V for Verb, and Adv
for Adverb. Trees and bracketings do not have to be labelled but the labelled ones are more
informative than the unlabelled ones.
Constituent analysis was suitable for resolving certain ambiguities, by showing different
constituent structures, e.g.: (old (men and women)) vs. ((old men) and women).

Generative Linguistics
However, there were lots of ambiguities which constituent analysis could not resolve. For
instance, The lamb is ready to eat. has two distinct meanings (is ambiguous), but the
American structuralists could give it only one analysis: ((The lamb) (is (ready (to eat)))).
Their analysis remained on the surface and could not disambiguate structures which were
different in the deep.
The growing dissatisfaction with the limitations of structuralist linguistics led to the
emergence of a radically new type of linguistic analysis towards the end of the 1950s. This
has become known as transformational-generative linguistics, or just generative linguistics
(= generative grammar), for short. This kind of analysis distinguishes two levels of syntactic
analysis: a surface structure or S structure(which was recognized by the structuralists, too)
and an underlying abstract deep structure or D-structure (which was not recognized by the
structuralists). Transformational-Generative grammar is transformational because it
explains surface structure as being derived from deep structure by a series of changes:
transformations.
In this framework, ambiguous sentences have identical surface structures but different deep
structures, according to the different meanings. For instance, the ambiguous sentence The
lamb is ready to eat (whose two meanings can be paraphrased as ‘The lamb can eat’ and
‘Somebody can eat the lamb’) is derived from two different deep structures. Synonymous
sentences like It rained yesterday. and Yesterday it rained., however, derive from one
common deep structure and differ only on the surface.
Transformational-generative grammar is generative, because it can generate (i.e. produce,
define and explicitly characterize) all and only the grammatical sentences of a language. This
means that (a) by applying the rules of the grammar, we always get a syntactically well-
formed sentence, (b) this kind of grammar generates all the well-formed sentences of a
language, i.e. not only those that have been uttered but also those that have not been uttered
but could be uttered, and are, thus, potential sentences of the language. The number of
possible grammatical sentences in any language at any one time is infinite, but the rules
which make this infinite variation possible are finite (otherwise the native speaker would not
be able to learn them).
The founder and most influential representative to this day of generative linguistics has been
the American linguist Noam Chomsky, whose works have found a great many followers all
over the world. Since its appearance the theory has been modified and remodified several
times and several new proposals have been made and are still being made by Chomsky
himself and by others.
As we saw above, the Bloomfieldians were uninterested in general theoretical questions,
emphasized the differences between individual languages, and thought that the main purpose
of linguistics was to describe individual languages. In contrast, Chomsky holds that
linguistics should be primarily concerned with Universal Grammar, i.e. with the principles
that are the properties of all human languages. One of these principles is structure-
dependence, which means that operations in a sentence apply to phrases and not just words,
i.e. these operations require a knowledge of the structural relationships of words rather than
just their linear sequence. For instance, when English speakers transform a declarative
sentence into a yes-or-no interrogative, the auxiliary they move is not simply “the second
word” of the declarative sentence, as a superficial observer might think on the basis of (4a),
but rather the word after the entire Noun Phrase that occupies the subject-slot of the
declarative sentence, as is shown in (4b).
(4) a. [NP John] will buy a car. _
Will [NP John] — buy a car?
b. [NP The man who has sold his house] will buy a car. _
Will [NP the man who has sold his house] — buy a car?
According to Chomsky, a generative grammar is a model for the native speaker’s intuitive
knowledge of the language (i.e. his internal grammar), a decisive part of which is Universal
Grammar and is genetically inherited. Chomsky calls the native speaker’s language-
knowledge competence (or – to use his more recent term – I-language). But the knowledge
of language, competence, has to be distinguished from the actual use of that knowledge in
real-life situations, i.e. from performance. Performance is the actual use of competence and
it involves individual and situational features, imperfections, errors, memory limitations, time
limitations on the length of sentences, life-span limitations on the number of sentences
actually produced by the individual, etc. Chomsky’s distinction between competence and
performance reminds us of Saussure’s distinction between langue and parole. But while
Chomsky uses the term performance in very much the same sense as Saussure used the term
parole, there is considerable difference between competence and langue. Saussure’s langue
was static: it was the system of linguistic signs. Chomsky’s competence is dynamic: it puts
the generation of sentences in the centre of attention. Another difference is that Saussure
thought of langue as being in the collective consciousness of a community. Chomsky thinks
of competence as knowledge whose basis is given to every normal human being by birth, in
the sense that its structure is related to the structure of the human mind and so the basis of
competence is a universal characteristic of the human species.
On the basis of their competence, native speakers can do several things:
• They can produce and understand an infinite number of new grammatical sentences in their
language.
• They can distinguish between grammatical and ungrammatical formations (He went to
London vs. *Went London he to).
• They can interpret elliptical sentences, i.e. sentences with missing elements (Peter is happy
but John isn’t).
• They can perceive ambiguity (The lamb is ready to eat).
• They can perceive synonymy (The duck crossed the road vs. The road was crossed by the
duck).
• They can idealize utterances, i.e. they can disregard the imperfections and idiosyncratic
features of performance and reconstruct the grammatical sentences which the utterances
realize (*? The thought of those poor children
were really … WAS really ... bothering me.).
The last point has a very important consequence: generative linguistics has an I language
approach to the study of language. Earlier, both Saussure and the American structuralists in
the first half of the 20th century were convinced that the way to la langue led through the
observation of la parole. In other words, linguistic analysis had to be based on a corpus of
data taken from the linguistic behaviour (actual language-use) of people, i.e. from parole or
performance. This can be called the E-language approach. By contrast, generative linguists
think that linguistics is concerned with far more than what can be found in a corpus. Thus,
even if we do use a corpus for linguistic work, we shall have to “idealize” the data, i.e. free
them from the imperfections and idiosyncrasies of performance. This is what native speakers
automatically do when they understand other native speakers’ utterances. They do so
intuitively, on the basis of their competence (or I language). But then the real task of
linguistics should be the study of the native speakers’ competence (and especially the part of
it which can be regarded as Universal Grammar). This is more important than the actual
utterances found in a corpus. Competence can be examined by asking questions about
intuitions. Consequently, the linguist has the right to use his own and other people’s intuitions
in linguistic analysis. And if the linguist is a native speaker of the language he examines, he
can ask and answer questions about his own intuitions. Examining one’s own intuitions
concerning language is a kind of introspection. In other words, generative linguists can base
their theories not (only) on empirical facts but on introspection and on native speakers’
intuitions. However, this does not mean that they give up objectivity because their theories
can be submitted to subsequent empirical verification. (Only their method is different from
the inductive method of the preceding decades: their method is deductive, proceeding from
theories to empirical facts.) But the focus of their attention is undoubtedly on I-language:
they are interested not so much in the empirical facts themselves as rather in the knowledge
that enables speakers to produce those empirical facts.
Since competence resides in the individual language-user’s mind and is a device of the
reasoning activity of human beings, it is a mental, psychological phenomenon. Consequently,
by studying what linguistic competence is and how it works, we are actually studying what
the mind is and how the mind works. If language-competence is part of the human mind, then
the study of this competence, i.e. linguistics, is part of the study of the mind, i.e. psychology.
In other words: Chomsky’s conclusion is that linguistics is a branch of cognitive
psychology. Generative linguistics, then, has extended the status of psycholinguistics from
being a mere branch of linguistics, to being the dominant branch of modern linguistics.