Language and Linguistics: Foundations Explained

1.1 Introduction

Human language, that unique characteristic of our species, has been of interest throughout history. The scientific study of human language is called linguistics. A linguist, then, is not someone who speaks many languages (although many linguists do); such individuals are polyglots. Instead, a linguist is a scientist who investigates human language in all its facets, its structure, its use, its history, its place in society.

The form and structure of the kinds of linguistic knowledge speakers possess is the concern of theoretical linguistics. This theory of grammar – the mental representation of linguistic knowledge – is what this textbook is about. But the field of linguistics is not limited to grammatical theory; it includes a large number of subfields, which is true of most sciences concerned with phenomena as complex as human language.

1.2 Panini to Chomsky and After

An interest like human language appears to have arisen when the human species evolved in the history of time. No culture has left records that do not reveal philosophical or practical concerns for this unique human characteristic. Different historical periods reveal different emphases and different goals, although both interests have existed in parallel.

Egyptian surgeons were concerned with clinical questions; an Egyptian papyrus, dated ca. 1700 BCE, includes medical descriptions of language disorders following brain injury. The philosophers of ancient Greece, on the other hand, argued and debated questions dealing with the origin and the nature of language. Plato, writing between 427 and 348 BCE, devoted his Cratylus Dialogue to linguistic issues of his day, and Aristotle was concerned with language from both rhetorical and philosophical points of view.

The Greeks and the Romans also wrote the grammar and discussed the sounds of language and the structures of words and sentences. This interest continued through the medieval period and the renaissance in an unbroken thread to the present period.

Linguistic scholarship, however, was not confined to Europe; in India, the Sanskrit language was the subject of detailed analysis as early as the twelfth century BCE. Panini’s Sanskrit grammar dated ca. 500 BCE is still considered to be one of the greatest scholarly linguistic achievements. In addition, Chinese and Arabic scholars have all contributed to our understanding of human language.

The major efforts of the linguists of the nineteenth century were devoted to historical and comparative studies. Ferdinand de Saussure (1857–1913), a Swiss linguist in this tradition, turned his attention instead to the structural principles of language rather than to how languages change and develop, and in so doing, became a major influence on twentieth-century linguistics.

In Europe and America, linguists turned to descriptive synchronic studies of languages and to the development of empirical methods for their analysis. Scholars from different disciplines and with different interests turned their attention to the many aspects of language and language use. American linguists in the first half of the century included the anthropologist Edward Sapir (1884–1939), interested in the languages of the Americas, language, and culture, and language in society, and Leonard Bloomfield (1887–1949), himself a historical and comparative linguist, as well as a major descriptive linguist who emerged as the most influential linguist in this period. Both Sapir and Bloomfield were also concerned with developing a general theory of language. Sapir was a ‘mentalist’ in that he believed that any viable linguistic theory must account for the mental representation of linguistic knowledge, its ‘psychological reality’; Bloomfield in his later years was a follower of behaviorism, which was the mainstream of psychological thought at the time, a view that precluded any concern for the mental representation of language and, in fact, for the mind itself.

In Europe, Roman Jakobson (1896–1982), one of the founders of the Prague School of Linguistics, came to America in 1941 and contributed substantially to new developments in the field. His collaboration with Morris Halle and Gunnar Fant led to a theory of Distinctive Features in phonology, and Halle has remained one of the leading phonologists of the last decades. In England, phoneticians like Daniel Jones (1881–1967) and Henry Sweet (1845–1912) (the prototype for G. B. Shaw’s Henry Higgins) have had a lasting influence on the study of the sound systems of language.

In 1957 with the publication of Syntactic Structures, Noam Chomsky ushered in the era of generative grammar, a theory which has been referred to as creating a scientific revolution. This theory of grammar has developed in-depth and breadth. It is concerned with the biological basis for the acquisition, representation, and use of human language and the universal principles which constrain the class of all languages. It seeks to construct a scientific theory that is explicit and explanatory.

The chapters that follow are based largely on the developments in linguistic theory that have occurred since the publication of Syntactic Structures in 1957 and Aspects of the Theory of Syntax in 1965. In subsequent years, Chomsky has continued to develop his theory in such major works as Remarks on Nominalization (1970), Conditions on Transformations (1973), Lectures on Government and Binding (1981), Barriers (1986), Principles and Parameters in Syntactic Theory (1981), and The Minimalist Program (1995).

1.3 Aims of Linguistic Theory

Three key questions were posed by Chomsky in 1986 which remain pivotal in linguistics today:

What constitutes knowledge of language? (Competence)

How is knowledge of language acquired? (Acquisition)

How is knowledge of language put to use? (Performance/language processing)

As stated above, this text will primarily concern the first question viewed in relation to the second. The development of language from infancy provides insights into the nature and structure of language itself and therefore is discussed in each part. Understanding language use (performance), the main tenet of psycholinguistic research, depends on our understanding of what is being put to use. We will discuss the distinction between linguistic knowledge (competence) and use (performance) below.

1.4 What Constitutes Knowledge of Language? Grammar as the Representation of Linguistic Competence

Knowledge of a language permits one to connect sounds (or gestures in sign languages) with meanings, that is, to understand a spoken or signed utterance, and to express our thoughts through speech or signs. Note that the sign languages of the deaf are basically the same as spoken languages, using a gestural/visual modality instead of the sound/aural perceptual modality of speech. Therefore, except where specifically referring to speech sounds, discussing the nature and characteristics of language should be interpreted as referring to both spoken and signed languages.

Linguistic knowledge, as represented in the speaker’s mind, is called grammar. Linguistic theory is concerned with revealing the nature of mental grammar, representing speakers’ knowledge of their language.

If one defines grammar as the mental representation of one’s linguistic knowledge, then a general theory of language is a theory of grammar. A grammar includes everything one knows about the structure of one’s language – its lexicon (the words or vocabulary in the mental dictionary), its morphology (the structure of words), its syntax (the structure of phrases and sentences and the constraints on well-formedness of sentences), its semantics (the meaning of words and sentences) and its phonetics and phonology (the sounds and the sound system or patterns). A theory of grammar specifies the nature of each of these components and the universal aspects of all grammar.

Each of these different kinds of linguistic knowledge constitutes a component of mental grammar. But what kind of knowledge is this? What do speakers know? First, it must be noted that we are not speaking of conscious knowledge. Most of us (before taking a course in linguistics) are totally unaware of the extent of our tacit unconscious knowledge of our language. Furthermore, we have no idea of the complexity of this knowledge. Some of this complex knowledge will be revealed in the chapters to come. As a way of introduction, however, we can illustrate the nature of this linguistic competence as represented in the components of our mental grammar.

1.4.1 The lexicon

Every speaker of a language has a dictionary or lexicon in their head, with all the words they know, like cat, witch, cauldron, Macbeth, jester, vocabulary, slay, betray, love, and hate. It has been estimated that the average person knows from 45,000 to 60,000 words; these must be stored in the mental lexicon.

1.4.2 Morphology

A speaker of a language also knows how words are structured, that, for example, in English, words are composed of bare roots like a witch, or roots with suffixes like witch-es or words with prefixes like dis-en-chant or words with prefixes and suffixes like dis-enchantment, and furthermore, they know that these parts of words must occur in a certain order. That is, to put a suffix like ment at the beginning of a word – *mentenchant – would make it unacceptable or ungrammatical. (Throughout the book, linguistically ill-formed – unacceptable, or ungrammatical – words or phrases or sentences will be preceded by an asterisk *.) Since speakers can distinguish between acceptable and unacceptable forms, they accept lover but reject *erlove, for example, then this is part of our grammatical knowledge, represented in our mental grammar.

1.4.3 Syntax

Part of our linguistic knowledge tells us what constitutes a well-formed string of words, how to put words together to form phrases and sentences. Moreover, we know when such strings of words are grammatical (well-formed) or ungrammatical (ill-formed), as in the difference between (1) and (2):

1. Lear had three daughters.

2. *Had three Lear daughters.

Note that grammaticality does not depend on our having heard the sentence before or whether it is true since (3)

3. Lear had only two daughters- is a grammatical sentence, but according to the Shakespeare tragedy, it is not a true sentence. And since Lear is a character in a play and does not nor did not exist as a King with three daughters, the acceptability of (1) does not depend on whether there is a referent in the real world for the information being conveyed. Our knowledge of syntax (and semantics) also accounts for the fact that we know that (4) is an ambiguous sentence (a sentence with more than one meaning).

4. Cordelia loved Lear more than Regan.

Although one meaning may come to mind first upon our hearing such a sentence, perhaps

4a. Cordelia loved Lear more than Regan loved Lear.

Speakers of English also know that (4b) may be the meaning of (4):

4b. Cordelia loved Lear more than Cordelia loved Regan.

The nature of syntax also accounts for the fact that there is an unlimited – infinite – set of sentences in any language. One cannot put a limit on the length of a sentence and thus cannot put a limit on the number of sentences.

5. Lear loved Cordelia.

6. Lear loved Cordelia and Cordelia loved Lear.

7. Lear, who had three daughters, loved Cordelia the most.

8. Lear, who loved Cordelia, became very angry with her when she would not tell him in words how much she loved him, and that made him cut her off without any lands or riches, which pleased Regan and Goneril but was very unfair because Cordelia really loved him the most and was unselfish and kind and all of this led to the terrible tragedy of King Lear.

We are able to embed sentences within sentences as shown in (9a–c):

9. a. Cordelia was Lear’s youngest daughter. b. She loved him. c. Cordelia, who was Lear’s youngest daughter, loved him.

We are also able to conjoin sentences as in (10):

10. Cordelia was Lear’s youngest daughter, and she loved him.

We know how to negate sentences and form questions.

11. Cordelia was not Lear’s youngest daughter.

Was Cordelia Lear’s youngest daughter?

We also know that in (13) him cannot refer to Lear, but in (14) him can refer to Lear or to someone else:

12. Lear loved him.

13. Regan told Lear she loved him.

Later chapters will deal with some of these questions and many more that have to do with our tacit knowledge of syntactic structure, which must also be represented in the mental grammar to account for our ability to make such judgments.

1.4.4 Semantics

Speakers also know quite a lot about what the expressions in their language mean or signify. This knowledge makes the patterns of sounds or gestures ‘symbolic’. A sentence is like a sort of acoustic or gestural picture – it represents something – though the way they manage to be representational is different from pictures. So, for example, we know that the spoken or written word Shakespeare can be used to name a person; and we know that the spoken or written phrase wrote plays signifies a property that some people but not others have, and we can put these two ideas together to recognize the meaning of a sentence like:

15. Shakespeare wrote plays.

The simple idea that the subject names something and the rest of the sentence tells us about that thing is not quite right, though. For example, this does not work for sentences.

16. No witches wrote plays.

Since sentence (16) is perfectly meaningful, no witches refer to any particular person. In chapters 7, 8, and 9, we will explain the meanings of sentences that handle both (15) and (16) equally well. Furthermore, we can understand sentences built up in all the ways that the syntax allows: we understand sentences that are embedded and coordinated; we understand negated sentences and questions. Our ability to recognize syntactic structures comes with this parallel ability to understand the infinite range of structures signifies. When we consider the meanings of expressions, we also notice that similar meanings can sometimes be conveyed in very different ways. For example, the following two sentences have very similar meanings:

17. Usually, a witch does not write plays.

18. Most witches do not write plays.

This kind of knowledge represents semantic knowledge in our mental grammar.

1.4.5 Phonetics and phonology

Speakers’ knowledge of their language also includes knowledge of the sounds and sound patterns which occur. We know what sounds are in the language and what sounds are not. Speakers of English know, unconsciously for the most part, that there are more than five vowel sounds in the language, as shown by the vowel sounds which differentiate the following words from each other: bit, beat, bet, bait, bat, boot, but, boat, bought, put, pot. We use five letters – a, e, i, o, u – to represent these different vowel sounds in our writing system. We see that there is no one-to-one mapping between alphabetic symbols and the sounds they represent.

Speakers of English also know what strings of sounds are words, are possible words, and are impossible words in their language. Thus, clasp occurs in most speakers’ mental dictionaries, while clisp or klisp does not (since no meaning is ‘attached’ to these sounds), but this nonsense form could become a word since it does not violate any constraints on sequences of sounds that are permitted in English. But *lkisp is not a possible word in English, nor is *ngisp since, in the first case, the sequence of the sounds lk cannot begin a word, and in the second case, ng, the sound which ends the word king, cannot begin a word. These are not ‘rules’ or ‘laws’ established by writers of textbooks but are constraints on the sound patterns of language which children learn when they acquire their language.

1.5 Mental Grammar, Universal Grammar, Descriptive Grammars, Teaching Grammars, and Prescriptive Grammars

Grammar, as viewed here, is different from the usual notion of grammar. When viewed as representing a speaker’s linguistic competence, grammar is a mental system, a cognitive part of the brain/mind, which, if it is one’s first native language, is acquired as a child without any specific instruction. The word grammar is often used solely about syntax. But we use it to refer to all aspects of linguistic competence. In addition to its use as referring to the mental system, when linguists describe this knowledge shared by a language community, the description is also called the grammar of the language. Of course, no two speakers of a language have identical grammars; some may know words that others do not, some may have some idiosyncratic rules or pronunciations. But since they can speak to each other and understand each other, there is a shared body of knowledge, which we call their mental grammar. Descriptive grammars are thus idealized forms of the mental grammars of all the speakers of a language community.

The grammars of all languages are constrained by universal ‘laws’ or ‘principles,’ a view which differs from that of many linguists in the pre-Chomsky period, some of whom held that languages could differ in innumerable ways. The more we look at the languages of the world, the more support there is for the position taken by Roger Bacon, a thirteenth-century philosopher, who wrote:

‘He that understands grammar in one language understands it in another as far as the essential properties of grammar are concerned. The fact that he can’t speak, nor comprehend, another language is due to the diversity of words and their various forms, but these are the accidental properties of grammar.’

There is much evidence to support this view, which today is based on recognizing that there is a biological basis for the human ability to acquire language. The child enters the world with an innate predisposition to acquire languages that adhere to these universal principles: a genetically determined mental system referred to as Universal Grammar or UG. This will be discussed further below.

While UG constrains the grammars of all human languages, there are the ‘accidental differences’ that constitute cross-linguistic variability. Using the theory of grammar, which specifies UG, linguists investigate and analyze specific languages and construct the descriptive grammars mentioned above. Thus, while UG may specify that the sound represented by the beginning the word thane (as in ‘Macbeth, the thane of Cawdor’) or thigh is a possible speech sound, the descriptive grammar of English will include this sound in its grammar but the grammar of French will not. Some languages permit the dropping of a pronominal subject; others do not. In some languages, one complex word may be equivalent to a whole sentence in another language. These differences will be revealed in the descriptive grammars of these languages. Descriptive grammars may also serve as the basis for teaching or pedagogical grammars used to teach someone a second language or a variation (dialect) of one’s native language.

Descriptive grammars aim at revealing mental grammar, which represents the knowledge a speaker of the language has. They do not attempt to prescribe what speakers’ grammar should be. While certain forms (or dialects) of a language may be preferred for social or political, or economic reasons, no specific dialect is linguistically superior to any other. The science of linguistics, therefore, has little interest in prescriptive grammars.

As discussed in chapter 2, it should also be noted that the majority of languages of the world have no written form. They are, however, as complex and rational as languages with a written orthography or alphabet. Speech (or sign) is primary, part of the natural endowment of the human species; writing systems are derived from spoken languages, which is why every normal human anywhere in the world who receives linguistic input as a child will learn the language of the environment but will not necessarily learn to read or write unless being specifically taught. Even deaf children deprived of linguistic input ‘invent’ their own gestural language, which develops and changes to conform to the constraints of UG. But such children do not construct written languages on their own.

1.6 How is Knowledge of Language Acquired? The Logical Problem of Child Language Acquisition

Young children, limited in so many respects, accomplish with apparent ease a remarkable cognitive feat. In just a few short years, without benefit of any direct instruction or correction, they develop a very complex and uniform cognitive system of linguistic knowledge, a grammar of the language being acquired. Just how children do, this is a central question that linguistic theory tries to answer. What makes the acquisition problem particularly intriguing is that we come to know vastly more about our language than provided evidence for in our linguistic environment. Consider, for example, the sentences in (19) and (20):

19. Polonius is eager to please.

20. Polonius is easy to please.

Superficially, these sentences are very similar. Only the adjective is different, easy in (20) and eager in (19). Yet, upon reflection, an English speaker will know that they differ in a way beyond the choice of adjective. In (19), the subject Polonius is the pleaser, while in (20), Polonius is the one being pleased, and someone else, unspecified in the sentence, is the pleaser. The sentence in (20) might be paraphrased as ‘It is easy to please Polonius,’ while the sentence in (19) has no such paraphrase. *It is eager to please John is not a good sentence in English. What we as speakers know about such sentences goes beyond what is exemplified in the superficial form. Similarly, the sentence in (21) has two meanings associated with it:

21. Visiting witches can be dangerous.

Speakers of English know that this sentence can mean (a) that it is dangerous to visit witches, or (b) witches visiting are dangerous. The ambiguity of such a sentence is not given in the superficial form of the sentence. Yet, it is something we know. Just as we know that the sentence in (22) is a grammatical question in English while the sentence in (23) is not, though (22) and (23) mean essentially the same thing.

22. What did Caesar drink nectar with?

23. *What did Caesar drink nectar and?

Other examples of what constitutes our linguistic knowledge were mentioned above. The point is that we know more about our language than meets the eye – or ear. How did we come to know so much about the structure and meaning of sentences in our language when what we hear are simply sequences of sounds? This problem of explaining the ease, rapidity, and uniformity of language development in the face of impoverished data is called the logical problem of language acquisition and was first posed in this form by Noam Chomsky (1955).

In the course of acquiring a language, children are exposed to only a finite set of utterances. Yet, they come to use and understand an infinite set of sentences, as discussed above. This has been referred to as the creative aspect of language use. This ‘creativity’ does not refer to the ability to write poetry or novels but rather the ability to produce and understand an unlimited set of new sentences never spoken or heard previously. The precise linguistic input children receive differs from child to child; no two children are exposed to exactly the same set of utterances. Yet, they all arrive at pretty much the same grammar. The input that children get is haphazard because caretakers do not talk to their children to illustrate a particular point of grammar. Yet, all children develop systematic knowledge of a language. Thus, despite the severe limitations and variation in the input children receive, and also in their personal circumstances, they all develop a rich and uniform system of linguistic knowledge. The knowledge attained goes beyond the input in various ways. How do we come to know as much as we do about our language if not from the linguistic environment?

In answer to the question of the logical problem of language acquisition, it has been proposed that much of what we know about our language is not learned from the input but is rather part of an innate endowment, which we referred to above as Universal Grammar (UG). UG specifies the form and functioning in general, hence principles that hold in all languages. On this view, the child’s mind does not approach language as a tabula rasa (a blank slate) to be written on by experience alone or armed only with general problem-solving skills such as imitation, memorization, analogy, or general induction. Rather, children are equipped with a set of specific expectations about linguistic structure and the principles which govern language. UG helps them overcome the limitations of the input and guides their grammatical development in particular ways. So children develop language rapidly and efficiently, that is, with relatively few errors, and despite the poverty of the stimulus (for example, the lack of negative evidence), because the basic form of language is given to them by human biology. Our common-sense understanding of language acquisition is that children just ‘pick up’ language. This seems to be close to the truth. It may be more appropriate to think in terms of language ‘growth’ rather than learning. This is not to say that there is no learning. Children must, of course, learn the lexicon of the particular language they are exposed to and other language-specific properties. As we will see throughout this book, languages such as English, French, Japanese, Swahili, and so on share many essential properties – those that derive from UG – but they also differ from each other in various respects, and the child must learn these differences based on experience. The best way to think about UG is as a template with gaps that must be filled in through experience with a particular language.

There is an intimate connection between linguistic theory and language acquisition. By analyzing the structure of individual languages, linguists try to determine which aspects of our linguistic knowledge are universal and hence, arguably, available to the child as part of UG, and which aspects are language-particular and hence to be learned based on linguistic input that the child receives. Thus, the study of particular grammar tells us something important about language development. At the same time, an understanding of the development of language in children can offer us further insight into the organization of human language.

1.7 How is Knowledge of Language Put to Use? Linguistic Performance

If grammar is the mental representation of linguistic competence, how does this differ from how we use this knowledge when we speak and comprehend what is said to us? The distinction between the representation of what we know in our minds and how we put this knowledge to use is not specific to language. For example, in performing music, we may ‘know by heart’ how to play the Moonlight Sonata, and we may play it perfectly from time to time. But we may also, at some performances, produce a few clinkers, make several mistakes, even forget a specific passage. The next time we try to play it, we may make no such mistakes, shows that the knowledge was there, but we just couldn’t get to it in our performance. That is, we know the sonata, but our performance reflects this knowledge in a non-perfect way.

This is also true of language. Although in principle, we can understand and produce an infinite set of sentences, obviously in our mortal lives, no speaker can do so. Although on hearing a sentence like (23) above, we know that there is something wrong with it, that it is ungrammatical, thus reflecting our linguistic competence, we may not state what part of our grammar is violated, showing that this is unconscious knowledge.

Differences between linguistic knowledge (competence) and linguistic performance are revealed, for example, through slips of the tongue. When the Reverend Spooner, whose errors gave birth to the term spoonerism, referred to Queen Victoria as that queer old dean instead of the intended dear old queen, he knew that she wasn’t a dean and wasn’t strange. Everyone makes errors, and often we catch ourselves doing it and correct the errors, showing that we know the correct form of the word, phrase, or sentence involved in the error.

Memory lapses sometimes prevent us from remembering the beginning of a sentence, producing errors like (24) in which a singular he is mistakenly produced instead of the plural they to agree with the plural subject Macbeth and Banquo.

24. Macbeth and Banquo, two generals of the King, rode to the castle, which he saw in the distance.

The relationship between grammar and linguistic performance is complex and is the major area of psycholinguistic research. This text will not discuss linguistic performance but rather the mental system accessed in speech and comprehension. However, given this distinction between competence and performance, how can linguists investigate competence, the nature of mental grammar?

1.8 ‘Doing’ Linguistics

There are a number of methods linguists use in ‘doing’ linguistics, in analyzing languages, in constructing grammars, and in developing a theory of grammar. Linguists use both naturalistic data and experimental data. Naturalistic data consist of actual speech written down or recorded and are often a useful source of positive evidence, that is, evidence that a particular type of sentence is (probably) grammatical. For example, suppose that in listening to recorded conversations or studying pages and pages of transcribed texts, one finds that the pronouns him and her always refer to someone other than the subject, in sentences like

25. Othello hates him.

26. Titania loved her.

While in sentences such as (27) and (28),

27. Othello hates himself.

28. Titania loved herself.

Himself and herself always refer to the subject; one can infer which pronouns