Introduction

This seminar paper is concerned with that part of our linguistic competence which includes our knowledge about individual words and their specific properties and which is called the lexicon. In the first part I want to deal with the question what is meant by the lexicon. My aim is it to examine the lexicon as a part of the grammar system, i.e. I want to analyse its nature, its status among the other components of the grammar and its relationship to them. Furthermore I want to point out the function of the lexicon within our linguistic competence. It shall become clear why we have to assume that there is something like the lexicon and why we need the information that is given in the lexicon in order to use the individual words of our vocabulary correctly.

In the second part of this seminar paper I want to analyse the information that is con-tained in the lexicon. The specific properties of individual words are partly semantic, pragmatic, morphological, syntactical and phonological properties. I want to deal mainly with the question which kind of semantic, morphological and syntactical information is given in a lexical entry. I will of course mention the other components of our lexical knowledge as well, but I want to focus on semantics, morphology and syntax.

---

1. The Lexicon as a Part of the Grammar

The grammar of a language can be defined as a finite set of rules, which determines which sentences, phrases or words are possible and which are not. We as native speakers have this set of rules internalized in our mind. It allows us (of course only under idealized circumstances) to understand and produce an infinite number of possible sentences of our native language. We even have a very subtle intuition by which we can analyse sentences and judge their grammatical correctness. This ability is called linguistic knowledge or competence. The aim of linguistics is it to find out and describe the rules that underlie a language. These rules can basically be subdivided into four different areas: Firstly there are the rules that determine the correct formation of sentence structures. They are examined in the syntax. Secondly there is the morphology, which deals with the rules of word forma-tion. Thirdly there is the phonology, which is concerned with describing the sound system of a language. And furthermore there is the semantics, which is the study of the meaning of words, phrases or sentences. These four areas - syntax, morphology, phonology and semantics - build the main components of the grammar. The purpose of studying a language is it to make out generalizations about the structures of that language and formulate general rules that are suitable to the linguistic competence of a native speaker.

But there are also certain areas of our linguistic knowledge which are not determined by general rules. The individual words that we know, use and understand and that make up our vocabulary have certain properties which cannot be generalized into rules, because they are specific to these words. Such specific properties of a word are, for example, its pronunciation and its meaning. The relationship between the pronunciation of a word and its meaning is (apart from a few onomatopoetic words like miaow or cuckoo) not regular, but completely arbitrary. The pronunciation and meaning of the word tree, for example, are not settled by rules, but they are specific to this word. They are, as I want to point out in the further course of this seminar paper, by far not the only specific properties of a word, but these examples shall show that our linguistic knowledge does not only consist of general rules. That means the sum of all rules that we as native speakers have internalized in our minds does not fully correspond to our linguistic knowledge. If we would define our linguistic knowledge merely as a set of rules, there would obviously be something missing.

Consequently we have to assume a further component of the grammar, where our knowledge about the specific and irregular properties of individual words is stored. This part of our linguistic knowledge is called the lexicon. Similar to a usual lexicon or dictionary in the form of a book we have a mental lexicon in our minds. But our mental lexicon is more than simply a list of all the words we know. In addition to the meaning and pronunciation of a word a lexical entry also contains a great deal of further information about morphological, syntactical, phonological, pragmatic and semantic properties, which are specific to that word or group of words. And we have to learn and to know these characteristic properties of each word of our vocabulary in order to use the words correctly. So the function of the lexicon is to provide us with the information we need to insert individual words in general rules.

---

2. The Information of the Lexicon

2.1 The Lexicon and Semantics

In this chapter I want to examine the linguistic knowledge that we have about individual words or phrases with regard to semantics. So the question I deal with is: What does our linguistic knowlegde that we have about the meaning of words look like? What comes first to oneīs mind when thinking about this question is probably the concrete meaning or denotation of a word; i.e. the relationship between the linguistic symbol (the phonetic form of a word) and the thought or idea this symbol stands for. This relationship is, as mentioned before, not determined by rules, but it is arbitrary. But this does not mean that everyone is free to change the meaning of a word. The meaning of words is settled by convention; they are so to speak "agreed" by the members of a speech community. Our knowledge about the meaning of words consists of pieces of information called semantic pro-perties. It can be split up into these semantic properties and vice versa they constitute the meaning of a word. For the word boy, for example, the semantic properties are "human", "male" and "young" and these pieces of information are contained in the mental lexicon of all speakers who know the meaning of boy. If speakers of English hear or read the word boy, in everybodyīs mind appears roughly the same picture. The individual "boy" in each mind will of course look different, but the thought is for everyone the same. If the members of a speech community would not share roughly the same concepts of meaning in their mental dictionaries, it would be obviously impossible to communicate with each other, because nobody would understand what the other one means.

The main semantic information given in a lexical entry is certainly the denotation of a word, but the lexicon is more than simply a list of words and their respective meaning; our semantic knowledge consists of more information. Apart from the denotation we also know, in the case of most content words, the connotation, i.e. the things that are not directly implied in the meaning of a word, but that are associated with this word. The word moon, for example, means first of all an astronomic body, which circles the earth. But it can also bear an additional secondary meaning, which is beyond its direct meaning; often things like "night", "coolness" or "love" are associated with the word moon. In addition to that we know a lot about the relations among words. Content words are not independent units on their own, but they are often semantically related to other words. This does of course not mean that each word is related to every single word of our lexicon, but there exist lexical or semantic relations among certain groups of words. The words elephant and carrot, for example, are hardly related in terms of their meaning. But our knowledge about these two words includes that elephant, as well as other words like horse, dog, cow, bird etc., is a hyponym of the superordinate animal and that the same is true for carrot, potato, cucumber, celery etc. in relationship to the word vegetable. So our knowledge about a word always comprises information about its relation to other words of our lexicon. Knowing the meaning of death, for example, includes knowing the meaning of life. And if we know the meaning of man, we also know that its opposite is woman. These two pairs of words are examples of complementary antonymy, which means that they are directly opposed to each other and being one excludes inevitably being the other one. Other types of antonyms are gradable antonyms like warm and cold and converse antonyms like parent and child. And again it does only make sense if we know both meanings and the way they are related. Further examples for lexical relations are synonymy and homonymy. Synonymy means that there are two different words that denote one and the same thing like couch and sofa or to close and to shut. So if we hear or read a word we, are mostly able to come up with two or three further words which have the same or a similar meaning. The word bat is an example of homonymy, because it bears two different meanings. Consequently there have to be two independent lexical entries for the word bat. A first one that says: A small animal which resembles a mouse and has wings to fly. And a second one that says: A piece of wood that is used for hitting a ball. A speaker of English has to be aware of this lexical ambiguity and decide from the context in which the word appears which of the two meanings is being referred to.

The meanings of individual words that we have stored in our lexicon can be combined into larger units of meaning like phrases, sentences or texts. Usually this happens by adding up the single meanings to a whole. If we combine the words the, old and tree, for example, we denote a certain object that has the semantic properties of old and tree. But this way of combining meaning does not work for idioms like to leave somebody in the lurch. I assume that there is a possibility of understanding this phrase literally, i.e. by deriving its meaning from the meanings of the individual words it consits of, but usually this phrase means to leave somebody in an unpleasant or difficult situation. So this phrase has a further idiomatic meaning, which cannot be concluded from the meanings of the single words. Consequently there has to be an independent lexical entry for this whole phrase, which contains this additional meaning. A further example of idiomaticity are certain phrasal verbs. The phrasal verb to fall down can be understood literally, because the two components keep their meaning when combined. The phrasal verb to turn down can be understood literally as well, but in addition to that it has an idiomatic meaning, which is to reject or refuse something and therefore it requires a seperate lexical entry. From that point of view idiomatic phrases have the same status as single words, because they have their own independent meaning (cf. Fromkin/Rodman 1993 : 121-138).

Finally I want to deal with a further aspect of our linguistic knowledge about words. This aspect is usually regarded as an independent component of the grammar and I cannot examine it in detail, but I think it is important to mention it. What I mean is our pragmatic knowledge about words, i.e. our knowledge about the properties or effects words may have in certain concrete speech acts. Words like to commence, to deem or conversant with, for example, appear quite formal, whereas other words that mean the same like to start, to consider or familiar with sound more natural and common. Some words on the other hand may appear informal or even derogatory or offensive. So according to the actual situation of our speech we have to choose the words that are appropriate to this situation. Depending on whether we have an important meeting with our boss or whether we have a talk with our friends in a pub, we certainly use a different style of language. Consequently our knowledge about words has to include how they appear when actually spoken. There has to be a mark in our lexical entries that helps us choose the right words for the respective situation.

2.2 The Lexicon and Morphology

Morphology is concerned with describing the processes of word formation. The question I want to deal with in this chapter is how these processes are related to the lexicon or how they result in the lexicon. On the one hand there are inflectional processes, i.e. inflectional affixes are added to the root of a word like in jump-s, jump-ed or jump-ing. This has the purpose of forming different grammatical forms of a word and expressing diverse grammatical properties like the tense, the number, the person, the mood etc. In the given examples the affixes indicate 3rd person singular present tense, past tense and present participle respectively. But the inflectional affixes cause only relative small changes in the properties of a word. Firstly there is a relative slight change in the meaning, because the inflectional affixes show us, for example, if there is only one or if there is more than one thing involved or they indicate in which tense something takes place and so on. And secondly they change the distribution of the word, i.e. they change the possible surroundings in which the word can appear. This becomes apparent by the fact that it is impossible to substitute different inflectional forms for each other without changing the meaning or the grammatical correctness of the sentence, as the following examples show:

(1) The horses jump over the fence.

(2) *The horses jumps over the fence.

(3) The horses jumped over the fence.

(4) *The horses jumping over the fence.

So in a way inflection changes the properties of a word. But these changes are not really specific, because inflection is fully productive. It is regular and it can always be applied, i.e. inflectional affixes like -ed, -ing or -s can be added to most verbs of the lexicon. Moreover the changes of the properties underlie a general rule. They are always to a hundred percent predictable. Adding an -s to a verb, for example, means always 3rd person singular present tense. If sentence number (2) is changed according to this rule, the result will be a correct sentence:

(5) The horse jumps over the fence.

As a conclusion it can be said that inflection changes in a regular and predcitable manner the grammatical form of a word. Actually it is still the same word. It keeps its grammatical category and its basic meaning. Inflection does not really produce new words, which form an independent lexical unit, i.e. a seperate lexical entry for each inflectional form of a word is not necessary. Usually we only need the roots or basic words, the several inflectional affixes and a general rule for their combination. An exception to this are however words that have irregular inflections. A noun like woman, for example, does not form its plural by adding an -s, but the plural of woman is women. And an irregular verb like to take does not form its tenses with the affix -ed, but the past tense and past participle forms are took and taken. Such cases are deviations from the rule and they are a specific property of a word. Consequently they require a seperate indication in the lexicon.

On the other hand there are the morphological processes of derivation and compound-ing. Derivation means that derivational affixes like -ment, -able, un-, -al etc. are added to roots as in think-able, develop-ment, un-known and nation-al. Compounding however is the combining of two roots like in home-work, book-shop or snow-man. What distinguishes derivation and compounding from inflection is the fact that there is a more significant change in the properties of derived or combined words. Firstly the meaning changes to a larger extent. The prefix un-, for example, can turn a word into its opposite as the example of un-known shows. And secondly the distribution of a word can be changed more significantly, because many derivational affixes are capable of changing the grammatical category of the word they are added to. The noun nation, for example, becomes an adjective when combined with -al and the consequence is of course a change in its distribution:

(6) The nation was upset.

(7) * The national was upset.

In contrast to inflection however these changes are less regular and less predictable. On the one hand it is true that there can be made out certain rules for derivation and compounding, but on the other hand these rules are not fully productive, i.e. they cannot be generalized. It is, for example, possible to derive the nouns owner, player, singer and writer from the corresponding verbs by adding the suffix -er, but this is not possible for the verb to steal; the word stealer is ungrammatical. And compared with inflection the semantic changes caused by derivation and compounding are also less regular and less pre-dictable. Adding an -s to a verb has always the same meaning, but adding the suffix -ize, for example, to a noun can cause different changes in meaning (compare to computerize, to womanize and to nationalize).

So the point I want to make is that derivation and compounding do not produce different grammatical forms of just one and the same word. On the contrary, the result of these processes are new and independent words with their own specific properties. Therefore compounds and words that are derived from other words require their own and independent lexical entry. This shows that the lexicon is not a fixed list of words. The area of derivation and compounding is an example of the creativity of language and according to this the lexicon can be enlarged by new words (cf. Huddleston 1984 : 25-32).

2.3 The Lexicon and Syntax

A further part of our linguistic competence is our abilitiy to combine the individual words of our vocabulary to syntactically well formed sentences. The syntax is concerned with finding out the rules that determine the structures of sentences. A possible approach to this is to classify the various types of phrases which a sentence may consist of and formulate rules for the internal structure of these phrases. The result are the following phrase structure rules:

(8)

S -> NP VP
VP -> V (NP) (PP)
NP -> (Art) (Adj) N (PP)
PP -> P NP

The letters on the left of the arrow denote the several phrasal categories (sentence, verb phrase, noun phrase and prepositional phrase) and the arrow means that the phrasal category on the left consists of the elements shown on the right side. The elements in brackets are optional, i.e. they may be a part of the respective phrase. So the obligatory components of a sentence, for example, are a VP and an NP. With the help of tree diagrams like the following it is possible to describe the hierarchy among the phrases of a sentence and the internal order of the words within the phrases:

(9)

At the upper level of a tree diagram are the phrasal categories (S, VP, NP, PP). These phrases however can be split up into lexical categories like noun (N), verb (V), article (Art), preposition (P) etc. At this point phrase structure rules and tree diagrams are generalizations about possible sentences of a language. They describe our knowledge of sentence structures. But they are quite general and certain difficulties arise when the concrete individual words from our mental dictionary are inserted into these general rules. The words of our vocabulary can be subdivided into different lexical or syntactical categories according to their distribution within a sentence. That means the different categories like nouns, verbs, prepositions etc. occupy certain fixed positions within a sentence or phrase. The fixed word order for an NP is, as mentioned above, (Art) (Adj) N. If we want to form a syntactically correct sentence, i.e. if we want to insert the words from our lexicon correctly into this general pattern, our knowledge about individual words has to include to which syntactical category they belong. The fact that we are able to distinguish different syntactical categories and to assign the words from our vocabulary to these categories together with our knowledge about the general phrase structure rules prevents us from forming wrong sentences like * Man the car bought a. Some words like surprise or cover can belong to two different categories. This of course requires two separate lexical entries.

So on the one hand the membership of syntactic classes is a specific property of a word, which we have to know in order to use the word correctly. But this is not the only knowledge we need, because words have more syntactic properties. The various syntactial categories themselves have to be subdivided again into different subcategories, because not each member of a category behaves syntactically in the same way. Verbs are not the only category that has to be subcategorized. There are also different types of nouns like proper nouns, countable nouns and uncountable nouns. But the distinctions in the syntactic behaviour of verbs are especially complicated, because there is a wide range of subclasses of verbs, which require different types and numbers of complements. Traditionally there is the distinction between intransitive, transitive and bitransitive verbs. Intransitive verbs like cry, sleep and scream do not require any complements; they can appear just on their own. Transitive verbs like buy, admire and hit take one and ditransitive verbs like give, offer and lend take two objects. Compare the following sentences:

(10) Susan screamed.
(11) Frank bought a book.
(12) Anna gave Mark the news paper.

But the distinction between intransitive, transitive and ditransitive verbs is a bit vague, because the various verbs require not only different numbers of complements, but also different types. In the examples above the complements are all noun phrases, but some tran-sitive verbs like lie or believe and some ditransitive verbs like put require a prepositional phrase as complement. And many verbs require a particular preposition; believe, for example, is always followed by an in-phrase, whereas rely is followed by on. Moreover there are other types of complements like adjective phrases or even sentences. And a lot of verbs take not only one type of complement, but there are often further alternative complements. Compare the following sentences:

(13) Joan lay on the bed.
(14) Tom doesnīt believe in God.
(15) Tom didnīt believe him.
(16) Tom didnīt believe that the story was true.
(17) Bill relied on Tom.
(18) Ben put the book on the table.
(19) He gave Frank the apple.
(20) He gave the apple to Frank.
(21) Robert smelled the smoke.
(22) The food smells very good.

So each verb takes a particular number and particular types of complements, i.e. each verb has its own characteristic complement structure. Instead of subdividing verbs into intransitive, transitive and ditransitive verbs it is more appropriate to say that each verbs has its own subcategorization frame. The subcategorization frame is a description of the complement structure of a verb. It includes in addition to the rather general information about the syntactical category of a verb information about the number and the types of complements that a verb requires. And verbs like believe, give and smell have more than one subcategorization frame. In the following there are some examples of subcategorization frames of verbs. Firstly there is an indication of the syntactical category and secondly the line in the square brackets is the postion where the word is inserted followed by the respective complements:

(17)

SCREAM: V; [ ___ ]
BELIEVE: V; [ ___ in-P]
BUY: V; [ ___ NP]
BELIEVE: V; [ ___NP]
GIVE: V; [ ___ NP NP]
BELIEVE: V; [ ___ S]
GIVE: V; [ ___ NP PP]
PUT: V; [ ___ NP PP]
LIE: V; [ ___ PP]
SMELL: V; [ ___ AP]
RELY: V; [ ___ on-P]
SMELL: V; [ ___ NP]

The subcategorization frame is a specific syntactic property of a verb, which is not regular or predictable and which distinguishes the various verbs from each other. It is a crucial part of our lexical knowledge about the syntactic behaviour of verbs, because it specifies the distribution of a verb, i.e. it determines the possible surroundings of a particular verb. We need this knowledge about the complement structure of a verb in order to insert the verb only in appropriate surroundings and prevent us from building sentences like the following (cf. Fromkin/Rodman 1993 : 92-113):

(18) * Joan lay the bed.
(19) * Bill doesnīt believe on God.
(20) * Ben put the book the table.
(21) * Susan cried the apple.
(22) * Robert admires.
(23) * The boss offered Tom.
(24) * He hit the ball the bat.

---

3. Conclusion

At this point of my seminar paper it should have been become clear that the lexicon is a term for our knowledge about the pronunciation and semantic, pragmatic, morphological and syntactical properties of words and that this knowledge is, apart from the general rules, a quite important part of our linguistic competence. Certain parts of our linguistic competence are according to mentalistic theories universal to all languages and they do not have to be learned, because they are innate to all human beings. The lexicon however does obviously not belong to a universal grammar, because its information are too specific. As native speakers we have to learn this knowledge during the process of language acquisition. That the information of the lexicon is learned and not innate becomes clear by the example of orthography or uncommon words. We only learn how to write a word at the age of six or seven when we have already learned the rest of our knowledge. And there are some words that children do not know, because they have not learnt them yet. If the lexicon were innate, this would not be the case.

Native speakers acquire the linguistic knowledge described as the lexicon amazingly quick and without any great effort. But everyone who ever tried to learn a foreign language knows how wide, complicated and subtle this knowledge is. The general rules of a foreign language can be learned relative easily, but it is very toilsome to learn and memorize the several specific properties of the many thousand words. Certain leanerīs dictionaries like the Oxford Advanced Learnerīs Dictionary or the Collinīs Cobuild Dictionary of Contemporary English serve as a help for foreign learners. Compared with learnerīs dictionaries dictionaries for native speakers manage with quite few information, because native speakers have most of the information already internalized. But foreign learnerīs on the other hand need a quite extensive dictionary, which functions as a replacement for a mental lexicon or helps them to acquire something that is roughly similar to a mental lexicon. Therefore if one compares the information given in an entry of a learnerīs dictionary with the information I have pointed out in this seminar paper, it will become clear that they resemble each other. One will certainly find the pronunciation, the meaning, an indication of the pragmatic effect, morphological irregularities and the subcategorization frame.

---

REFERENCES

Fromkin, Victoria and Robert Rodman, 1993. 5th ed. An Introduction to Language. New York: Harcourt Brace Jovanowitch.

Huddleston, Rodney, 1984. Introduction to the Grammar of English. London: C.U.P.

Glück, Helmut (ed.), 1993. Metzler-Lexikon Sprache. Stuttgart: Metzler.

OīGrady, William, Michael Dobrovolsky and Mark Aronoff, 1993. 2nd ed. Contemporary Linguistics: An Introduction. New York: St. Martinīs Press.

---