SEMANTICS

Chapter I
INTRODUCTION TO SEMANTICS

Why study semantics? Semantics (as the study of meaning) is central to the study of communication and as communication becomes more and more a crucial factor in social organization, the need to understand it becomes more and more pressing. Semantics is also at the centre of the study of the human mind – thought processes, cognition, conceptualization – all these are intricately bound up with the way in which we classify and convey our experience of the world through language.
Because it is, in these two ways, a focal point in man’s study of man, semantics has been the meeting place of various cross-currents of thinking and various disciplines of study. Philosophy, psychology, and linguistics all claim a deep interest in the subject. Semantics has often seemed baffling because there are many different approaches to it, and the ways in which they are related to one another are rarely clear, even to writers on the subject. (Leech 1990: IX).
Semantics is a branch of linguistics, which is the study of language; it is an area of study interacting with those of syntax and phonology. A person’s linguistic abilities are based on knowledge that they have. One of the insights of modern linguistics is that speakers of a language have different types of linguistic knowledge, including how to pronounce words, how to construct sentences, and about the meaning of individual words and sentences. To reflect this, linguistic description has different levels of analysis. So – phonology is the study of what sounds combine to form words; syntax is the study of how words can be combined into sentences; and semantics is the study of the meanings of words and sentences.

1. A Short History of Semantics
It has often been pointed out, and for obvious reasons, that semantics is the youngest branch of linguistics (Ullmann 1962, Greimas 1962). Yet, interest in what we call today “problems of semantics” was quite alive already in ancient times. In ancient Greece, philosophers spent much time debating the problem of the way in which words acquired their meaning. The question why is a thing called by a given name, was answered in two different ways.
Some of them believed that the names of things were arrived at naturally, physei, that they were somehow conditioned by the natural properties of things themselves. They took great pains to explain for instance that a letter like “rho” seems apt to express motion since the tongue moves rapidly in its production. Hence its occurence in such words as rhoein (“to flow”), while other sounds such as /s, f, ks/, which require greater breath effort in production, are apt for such names as psychron (“shivering”) or kseon (“shaking”), etc. The obvious inadvertencies of such correlations did not discourage philosophers from believing that it is the physical nature of the sounds of a name that can tell us something about its meaning.
Other philosophers held the opposite view, namely that names are given to things arbitrarily through convention, thesei. The physei-thesei controversy or physis-nomos controversy is amply discussed in Plato’s dialogue Cratylus. In the dialogue, Cratylus appears to be a part of the physei theory of name acquistion, while Hermogenes defends the opposite, nomos or their point of view. The two positions are then debated by Socrates in his usual manner. In an attempt to mediate between the two discussants he points out first of all that there are two types of names. Some are compound names which are divisible into smaller constituent element and accordingly, analyzable into the meaning of these constituent elements: Poseidon derives his name from posi (“for the feet”) and desmos (“fetter”) since it was believed that it was difficult for the sea god to walk in the water.
The words, in themselves, Socrates points out, give us no clue as to their “natural” meaning, except for the nature of their sounds. Certain qualities are attributed to certain types of sounds and then the meaning of words is analyzed in terms of the qualities of the sounds they are made of. When faced with abundant examples which run counter the apriori hypothesis: finding a “l” sound (“lambda”) “characteristic of liquid movements” in the word sklerotes (“hardness”) for instance, he concludes, in true socratic fashion, that “we must admit that both convention and usage contribute to the manifestation of what we have in mind when we speak”.
In two other dialogues, Theatetus and Sophists, Plato dealt with other problems such as the relation between thought language, and the outside world. In fact, Plato opened the way for the analysis of the sentence in terms which are parly linguistic and partly pertaining to logic. He was dealing therefore with matters pertaining to syntactic semantics, the meaning of utterrances, rather than the meaning of individual words.
Aristotle’s works (Organon as well as Rhetoric and Poetics) represent the next major contribution of antiquity to language study in general and semantics in particular. His general approach to language was that of a logician, in the sense that he was interested in what there is to know how men know it, and how they express it in langugage (Dinneen, 1967: 70) and it is through this perspective that his contribution to linguistics should be assessed.
In the field of semantics proper, he identified a level of language analysis – the lexical one – the main purpose of which was to study the meaning of words either in isolation or in syntactic constructions. He deepened the discussion of the polysemy, antonymy, synonymy and homony and developed a full-fledged theory of metaphor.
The contribution of stoic philosophy to semantics is related to their discussion of the nature of linguistic sign. In fact, as it was pointed out (Jakobson, 1965: 21, Stati 1971: 182, etc.) centuries ahead of Ferdinand de Saussure, the theory of the Janus-like nature of the linguistic sign – semeion – is an entity resulting from the relationship obtaining between the signifier – semainon – (i.e. the sound or graphic aspect of the word), the signified – semainomenon (i.e. the notion) and the object thus named – tynkhanon -, a very clear distinction, therefore, between reference and meaning as postulated much later by Ogden and Richards in the famous “triangle” that goes by their name.
Etymology was also much debated in antiquity; but the explanations given to changes in the meaning and form of words were marred on the one hand by their belief that semantic evolution was always unidirectional, from a supposedly “correct” initial meaning, to their corruption, and, on the other hand, by their disregard of phonetic laws (Stati, 1971: 182).
During the Middle Ages, it is worth mentioning in the field of linguistics and semantics the activity of the “Modistae” the group of philosophers so named because of their writings On the Modes of Signification. These writings were highly speculative grammars in wich semantic considerations held an important position. The “Modistae” adopted the “thesei” point of view in the “physei-thesei” controversy and their efforts were directed towards pointing out the “modi intelligendi”, the ways in which we can know things, and the “modi significandi”, the various ways of signifying them (Dinneen, 1967: 143).
It may be concluded that throughout antiquity and the Middle Ages, and actually until the 19th century almost everything that came to be known about meaning in languages was the result of philosophic speculation and logical reasoning. Philosophy and logic were the two important sciences which left their strong impact on the study of linguistic meaning.
It was only during the 19th century that semantics came into being as an independent branch of linguistics as a science in its own right. The first words which confined themselves to the study of semantic problems as we understand them today, date as far back as the beginning of the last century.
In his lectures as Halle University, the German linguist Ch. C. Reisig was the first to formulate the object of study of the new science of meaning which he called semasiology. He conceived the new linguistic branch of study as a historical science studying the principles governing the evolution of meaning.
Towards the end of the century (1897), M. Bréal published an important book Essay de sémantique which was soon translated into English and found an immediate echo in France as well as in other countries of Europe. In many ways it marks the birthday of semantics as a modern linguistic discipline. Bréal did not only provide the name for the new science, which became general in use, but also circumscribed more clearly its subject-matter.
The theoretical sources of semantic linguistics outlined by Bréal are, again, classical logic and rethorics, to which the insights of an upcoming science, namely, psychology are added. In following the various changes in the meaning of words, interest is focused on identifying certain general laws governing these changes. Some of these laws are arrived at by the recourse to the categories of logic: extension of meaning, narrowing of meaning, transfer of meaning, while others are due to a psychological approach, degradation of meaning and the reverse process of elevation of meaning.
Alongside these theoretical endeavours to “modernize” semantics as the youngest branch of linguistics, the study of meaning was considerably enhanced by the writing of dictionaries, both monolingual and bilingual. Lexicographic practice found extensive evidence for the categories and principles used in the study of meaning from antiquity to the more modern approaches of this science: polysemy, synonymy, homonymy, antonymy, as well as for the laws of semantic change mentioned above.
The study of language meaning has a long tradition in Romania. Stati mentioned (1971: 184) Dimitrie Cantemir’s contribution to the discussion of the difference between categorematic and syncategorematic words so dear to the medieval scholastics.
Lexicography attained remarkably high standards due mainly to B. P. Hasdeu. His Magnum Etymologicum Romaniae ranks with the other great lexicographic works of the time.
In 1887, ten years ahead of M. Bréal, Lazar Saineanu published a remarkable book entitled Incercare asupra semasiologiei limbei romane. Studii istorice despre tranzitiunea sensurilor. This constitutes one of the first works on semantics to have appeared anywhere. Saineanu makes ample use of the contributions of psychology in his attempts at identifying the semantic associations established among words and the “logical laws and affinities” governing the evolution of words in particular and of language in general.
Although it doesn’t contain an explicit theory of semantics, the posthumous publication of Ferdinand de Saussure’s Cours de linguistique générale 1916, owing to the revolutionary character of the ideas on the study of language it contained, determined an interest for structure in the field of semantics as well.
Within this process of development of the young linguistic discipline, the 1921-1931 decade has a particular significance. It is marked by the publication of three important books: Jost Trier, Der Deutsche Wortschatz im Sinnbezink des Verstandes (1931), G. Stern, Meaning and Change of Meaning (1931) and C. K. Ogden and J. A. Richards: The Meaning of Meaning (1923).
Jost Trier’s book as well as his other studies which are visibly influenced by W. von Humbold’s ideas on language, represents an attempt to approach some of the Saussurean principles to semantics. Analyzing the meaning of a set of lexical elements related to one another by their content, and thus belonging to a semantic “field”, Trier reached the conclusion that they were structurally organized within this field, in such a manner that the significative value of each element was determined by the position which it occupied within the respective field. For the first time, therefore, words were no longer approached in isolation, but analyzed in terms of their position within a larger ensemble – the semantic field – which in turn, is integrated, together with other fields, into an ever larger one. The process of subsequent integrations continues until the entire lexicon is covered. The lexicon therefore is envisaged as a huge mosaic with no piece missing.
Gustav Stern’s work is an ambitious attempt at examining the component factors of meaning and of determining, on this ground, the causes and directions of changes of meaning. Using scientific advances psychology (particularly Wundt’s psychlogy) Stern postulates several classifications and principles which no linguist could possibly neglect.
As regards Ogden and Richard’s book, its very title The Meaning of Meaning is suggestive of its content. The book deals for the most part with the different accepted definitions of the word “meaning”, not only in linguistics, but in other disciplines as well, identifying no less than twenty-four such definitions. The overt endeavour of the authors is to confine semantic preoccupations to linguistic problems exclusively. The two authors have the merit of having postulated the triadic relational theory of meaning – graphically represented by the triangle that bears their names.
A short supplement appended to the book: The Problem of Meaning in Primitive Languages due to an anthropologist, B. Malinowski, was highly instrumental in the development of a new “contextual” theory of meaning advocated by the British school of linguistics headed by J. R. Firth.
The following decades, more specifically the period 1930-1950 is known as a period of crisis in semantics. Meaning was all but completely ignored in linguistics particularly as an effect of the position adopted by L. Bloomfield, who considered that the study of meaning was outside the scope of linguistics proper. Its study falls rather within the boundaries of other sciences such as chemistry, physics, etc., and more especially psychology, sociology or anthropology. The somewhat more conciliatory positions which, without denying the role of meaning in language nevertheless alloted it but a marginal place within the study of language (Hockett, 1958), was not able to put an end to this period of crisis.
Reference to semantics was only made in extremis, when the various linguistic theories were not able to integrate the complexity of linguistic events within a unitary system. Hence the widespread idea of viewing semantics as a “refuge”, as a vast container in which all language facts that were difficult to formalize could be disposed of.
The picture of the development of semantics throughout this period would be incomplete, were it not to comprise the valuable accumulation of data regarding meaning, all due to the pursuing of tradition methods and primarily to lexicographic practice.
If we view the situation from a broader perspective, it becomes evident that the so-called “crisis” of semantics, actually referred to the crisis of this linguistic discipline only from a structuralist standpoint, more specifically from the point of view of American descriptivism. On the other hand, however, it is also salient that the renovating tendencies, as inaugurated by different linguistic schools, did not incorporate the semantic domain until very late. It was only in the last years of the sixties that the organized attacks of the modern linguistic schools of different orientations was launched upon the vast domain of linguistic meaning.
At present meaning has ceased to be an “anathema” for linguistics. Moreover, the various linguistic theories are unanimous in admitting that no language description can be regarded as being complete without including facts of meaning in its analysis.
A specific feature of modern research in linguistics is the ever growing interest in problems of meaning. Judging by the great number of published works, by the extensive number of semantic theories which have been postulated, of which some are complementary, while some other are directly opposed, we are witnessing a period of feverish research, of effervescence, which cannot but lead to progress in semantics.
An important development in the direction of a psycholinguistic approach to meaning is Lakoff’s investigation of the metaphorical basis of meaning (Lakoff and Johnson 1980). This approach draw on Elinor Rosch’s notion of protype, and adopt the view opposed to that of Chomsky, that meaning cannot be easily separated from the more general cognitive functions of the mind.
G. Leech considers that the developments which will bring most rewards in the future will be those which bring into a harmonious synthesis the insights provided by the three disciplines which claim the most direct and general interest in meaning: those of linguistics, philosophy and psychology.

2. Definition and Object of Semantics

In linguistic terminology the word semantics is used to designate the science of word-meaning. The term, however, has acquired a number of senses in contemporary science. Also, a number of other terms have been proposed to cover the same area of study, namely the study of meaning. As to meaning itself, the term has a variety of uses in the metalanguage of several sciences such as logic, psychology, linguistics, and more recently semiotics.
All these factors render it necessary to discuss on the one hand the terminology used in the study of meaning and on the other hand, the main concerns of the science devoted to the study of meaning.
One particular meaning of the term semantics is used to designate a new science, General Semantics, the psychological and pedagogical doctrine founded by Alfred Korzybsky (1933) under the influence of contemporary neo-positivism. Starting from the supposed exercise upon man’s behaviour, General semantics aims at correcting the “inconsistencies” of natural language as well as their tendency to “simplify” the complex nature of reality.
A clearer definition of the meaning (or meanings) of a word is said to contribute to removing the “dogmatism” and “rigidity” of language and to make up for the lack of emotional balance among people which is ultimately due to language. This school of thought holds that the study of communicative process can be a powerful force for good in the resolution of human conflict, whether on an individual, local, or international scale. This is a rather naïve point of view concerning the causes of conflicts (G. Leech 1990: XI). Yet, certain aspects of the relationship between linguistic signs and their users – speakers and listeners alike – have, of course, to be analyzed given their relevance for the meaning of the respective signs.
Also, that there is a dialectic interdependence between language and thought in the sense that language does not serve merely to express thought, but takes an active part in the very moulding of thought, is beyond any doubt.
On the whole, however the extreme position adopted by general semanticists as evidenced by such formulations as “the tyranny of words”, “the power of language”, “man at the mercy of language”, etc. has brought this “science” to the point of ridicule, despite the efforts of genuine scholars such as Hayakawa and others to uphold it.
In the more general science of semiotics, the term semantics is used in two senses:
(a) theoretical (pure) semantics, which aims at formulating an abstract theory of meaning in the process of cognition, and therefore belongs to logic, more precisely to symbolic logic;
(b) empirical (linguistic) semantics, which studies meaning in natural languages, that is the relationship between linguistic signs and their meaning. Obviously, of the two types of semantics, it is empirical semantics that falls within the scope of linguistics.
The most commonly agreed-upon definition of semantics remains the one given by Bréal as “the science of the meanings of words and of the changes in their meaning”. With this definition, semantics is included under lexicology, the more general science of words, being its most important branch.
The result of research in the field of word-meaning usually takes the form of dictionaries of all kinds, which is the proper object of the study of lexicography.
The term semasiology is sometimes used instead of semantics, with exactly the same meaning. However since this term is also used in opposition to onomasiology it is probably better to keep it for this more restricted usage. Semasiology stands for the study of meaning starting from the “signifiant” (the acoustic image) of a sign and examining the possible “signifiés” attached to it. Onomasiology accounts for the opposite direction of study, namely from a “signifié” to the various “signifiants” that may stand for it.
Since de Saussure, the idea that any linguistic form is made up of two aspects – a material one and an ideal one -, the lingistic sign being indestructible union between a signifiant and a signifié, between an expression and a content. In the light of these concepts, the definition of semantics as the science of meaning of words and of the changes in meaning, appears to be rather confined. The definition certainly needs to be extended so as to include the entire level of the content of language. As Hjelmslev pointed out, there should be a science whose object of study should be the content of language and proposed to call it plerematics. Nevertheless all the glossematicians, including Hjelmslev continued to use the older term – semantics in their works.
E. Prieto (1964) calls the science of the content of language noology (from Greek noos – “mind”) but the term has failed to gain currency.
Obviously, a distinction should be made between lexosemantics, which studies lexical meaning proper in the traditional terminology and morphosemantics, which studies the grammatical aspect of word-meaning.
With the advent of generative grammar emphasis was switched from the meaning of words to the meaning of sentences. Semantic analysis will accordingly be required to explain how sentences are understood by the speakers of language. Also, the task of semantic analysis is to explain the relations existing among sentences, why certain sentences are anomalous, although grammatically correct, why other sentences are semantically ambiguous, since they admit of several interpretations, why other sentences are synonymous or paraphrases of each other, etc.
Of course, much of the information required to give an answer to these questions is carried by the lexical items themselves, and generative semantics does include a representation of the meaning of lexical elements, but a total interpretation of a sentence depends on its syntactic structure as well, more particularly on how these meanings of words are woven into syntactic structure in order to allow for the correct interpretation of sentences and to relate them to objective reality. In the case of generative semantics it is obvious that we can speak of syntactic semantics, which includes a much wider area of study that lexical semantics.

3. Semantics and Semiotics

When the Stoics identified the sing as the constant relationship between the signifier and the signified they actually had in mind any kind of signs not just linguistic ones. They postulated a new science of signs, a science for which a term already existed in Greek: sêmeiotikê. It is however, only very recently, despite repeated attempts by foresighted scientists, that semiotics become a science in its own right.
A first, and very clear presentation of semiotics is it to be found in this extensive quotation from John Locke’s An Essay Concerning Human Understanding. In the chapter on the “division of the sciences”, Locke mentions “the third branch (which) may be called semiotic, or the doctrine of signs… the business whereof is to consider the nature of signs the mind makes use of for the understanding of things, or conveying its knowledge to others. For, since the things the mind contemplates are none of them, beside itself, present to the understanding, it is necessary that something else, a sign or representation of the thing it considers, should be present to it” (Locke, 1964: 309).
Later, in the 19th century, the American philosopher Charles Sanders Peirce devoted a life time work, which unfortunately remained unheeded for a long time, to the study of signs, to setting up semiotics as a science, “as the doctrine of the essential nature and fundamental varieties of possible «semiosis»”. (R. Jakobson, 1965: 22). Ferdinand de Saussure too, probably quite independently from Peirce, but undoubtedly inspired by the same Greek philosophers’ speculations on language, suggested that linguistics should be regarded as just one branch of a more general science of sign systems which he called semiology. In other words he saw no basic difference between language signs and any other kinds of sings all of them interpretable by reference to the same general science of signs.
Peirce distinguished three main types of signs according to the nature of the relationship between the two inseparable aspects of a sign: the signans (the material suport of the sign, its concrete manifestation) and the signatum (the thing signified):
(i) Icons in which the relationship between the signans and the signatum is one of the similarity.
The signans of an iconic type of sign, resembles in shape its signatum. Drawings, photographs, etc., are examples of iconic signs. Yet, phisical similarity does not imply true copying or reflection of the signatum by the signans. Peirce distinguished two subclasses of icons-images and diagrams. In the case of the latter, it is obvious that the “similarity” is hardly “physical” at all. In a diagram of the rate of population or industrial production growth, for instance, convention plays a very important part.
(ii) Indexes, in which the relationship between the signans and the signatum is the result of a constant association based on physical contiguity not on similarity. The signans does not resemble the signatum to indicate it. Thus smoke is an index for fire, gathering clouds indicate a coming rain, high temperature is an index for illness, footprints are indexes for the presence of animals, etc.
(iii) Symbols, in which the relationship between the signans and the signatum is entirely conventional. There is no similarity or physical contiguity between the two. The signans and signatum are bound by convention; their relationship is an arbitrary one. Language signs are essentially symbolic in nature. Ferdinand de Saussure clearly specified absolute arbitrariness as “the proper condition of the verbal sign”.
The act of semiosis may be both motivated and conventional. If semiosis is motivated, than motivation is achieved either by contiguity or by similarity.
Any system of signs endowed with homogeneous significations forms a language; and any language should be conceived of as a mixture of signs.
Another aspect revealed by semiotics which presents a particular importance for semantics is the understanding of the semiotic act as an institutional one. Language itself, can be regarded as an institution (Firth, 1957), as a complex form of human behaviour governed by signs. This understanding of language opens the way for a new, intentional theory of meaning. Meaning is achieved therefore either by convention or by intention.

Bibliography:
1. Chiţoran, Dumitru. 1973. Elements of English Structural Semantics. Bucureşti: E.D.P.
2. Leech, G. 1990. Semantics. The Study of Meaning. London: Penguin Books.
3. Saeed, J., I. 1997. Semantics. Dublin: Blackwell Publishers.
TOPICS FOR DISCUSSION
1. Define semantics and its object.
2. The physei-thesei controversy.
3. Types of signs.

Chapter II
THE PROBLEM OF MEANING

I. The concept of meaning 1.a bipolar relation
2. a triadic relation – A. referential approach
– B. conceptual approach
3. Heger’s view.
II. Dimensions of meaning 1. dimensions of meaning
2. types of meaning in Leech’ s conception.
I.1. Any progress in semantics is conditioned by a clearer understanding of meaning, as the object of its analysis. Numberless definitions of language meaning have been postulated, some complementary in nature, some opposed. A linguistic account of meaning would still be very difficult to give because of the plurality of levels at which meaning can be discussed- the word level, the phrase level, the sentence level.
Even if the morpheme is the minimum unit of language endowed with meaning, it is the word, the next higher unit that traditional lexicology has selected as its object of study and to clearly understand the factors involved in meaning, it’s necessary to begin with an account of meaning at word level.
The concept of meaning, defined by F. de Saussure, was first regarded as a bipolar relation between the two interdependent sides of a linguistic sign-significans ‘expression’ and significatum ‘cont3ent’ and this is true for any sign, no matter to what semiotic system it belongs.
2. Ogden and Richards have pointed out in 1923 that at least three factors are involved in any symbolic act- the symbol itself ‘the material aspect of the linguistic sign, be it phonic or graphic’; the thought/reference ‘the mental content that accompanies the occurrence of the symbol in the minds of both the speaker and the listener’; the object itself/ the referent ‘the object in the real world designated by the symbol’.
The triadic concept of meaning was represented by Ogden and Richards in the form of a triangle.

While the relation symbol- reference and reference- referent are direct and causal ones in the sense that the symbol expresses or symbolises the reference which, in turn refers to the referent, the relation symbol- object or referent is an imputed, indirect one.
Of the two sides of the triangle only the right-hand one can be left out – tentatively and temporarily- in a linguistic account of meaning. The relationship between thought and the outside world of objects and phenomena is of interest primarily to psychologists and philosophers, linguists directing their attention towards the other two sides. (Chiţoran, 1973: 30).
Depending on what it is understood by meaning, we can distinguish two main semantic theories:
– the referential / denotational approach-meaning is the action of putting words into relationship with the world;
– the representational /conceptual approach-meaning is the notion, the concept or the mental image of the object or situation in reality as reflected in man’s mind.
The two basic types of meaning were first mentioned by S. Stati in 1971- referential definitions which analyse meaning in terms of the relation symbol- object /referent; conceptual definitions which regard the relation symbol- thought/reference.
A. Denotational /Referential Theories of Meaning.
Before describing the characteristics of these theories, a clarification of the terms used is necessary. All languages allow speakers to describe or model aspects of what they perceive. In semantics the action of picking out or identifying individuals/ locations with words is called referring/denoting. To some linguists the two terms, denote and refer are synonymous. J. Saeed (1997: 23) gives two examples of proper names whose corresponding referents are easily recognizable
e. g. I saw Michael Jackson on TV last night.
We have just flown back from Paris.
The underlined words refer to/denote the famous singer, respectively the capital of France, even if in some contexts they may be used to designate a person different from the singer, or a locality other than the capital of France.
To John Lyons the terms denote and refer are not synonymous. The former is used to express the relationship linguistic expression- world, whereas the latter is used for the action of a speaker in picking out entities in the world. In the example
A sparrow flew into the room.
A sparrow and the room are NPs that refer to things in the world.; room, sparrow denote classes of items. In conclusion, referring is what speakers do and denoting is a property of words. Denotation is a stable relationship in a language, it doesn’t depend on anyone’s use of the word unlike the action of referring.
Returning to the problem of theories of meaning, they are called referential/ denotational when their basic premise is that we can give the meaning of words and sentences by showing how they relate to situations- proper names denote individuals, nouns denote entities or sets of individuals, verbs denote actions, adverbs denote properties of actions, adjectives denote properties of individuals-.In case of sentences, they denote situations and events. The difference in meaning between a sentence and its negative counterpart arises from the fact that they describe two situations
e. g. There is a book on the shelf.
There isn’t a book on the shelf.
Referential theories consider meaning to be something outside the world itself, an extra-linguistic entity. This means reducing the linguistic sign, i. e. the word to its material aspect, be it phonic or graphic.
The impossibility of equating meaning with the object denoted by a given word can be explained considering three major reasons
a. the identity meaning-object would leave meaning to a large extent undefined because not all the characteristic traits of an object as an extra- linguistic reality are identical with the distinctive features of lexical meaning;
b. not all words have a referent in the outside world; there are:
– non- referring expressions so, very, maybe, if, not, etc.
– referring expressions used generically:
e. g. A murder is a serious felony.
– words like nouns, pronouns with variable reference depending on the context:
e. g. The president decides on the foreign policy.
She didn’t know what to say.
– words which have no corresponding object in the real world in general or at a certain moment:
e. g. The unicorn is a mythical animal.
She wants to make a cake this evening.
– different expressions/words that can be used for the same referent, the meaning reflecting the perspective from which the referent is viewed
e. g. The morning star is the same thing as the evening star.
The president of the USA/ George Bush/ Barbara Bush’s husband was to deliver a speech.
Besides the referential differences between expressions, we can make useful distinctions among the things referred to by expressions-referent = thing picked out by uttering the expression in a particular context; extension of an expression = set of things which could possibly be the referent of that expression. In Lyon’s terminology the relationship between an expression and its extension is called denotation.(Saeed 1997: 27)
A distinction currently made by modern linguists is that between the denotation of a word and the connotations associated with it. For most linguists, denotation represents the cognitive or communicative aspect of meaning (Schaff 1965), while connotation stands for the emotional overtones a speaker usually associates with each individual use of a word. Denotative meaning accounts for the relationship between the linguistic sign and its denotatum. But one shouldn’t equate denotation with the denotatum.What is the denotation of a word which has no denotatum.
As far as the attitude of the speaker is concerned, denotation is regarded as neutral, since its function is simply to convey the informational load carried by a word. The connotative aspects of meaning are highly subjective, springing from personal experiences, which a speaker has had of a given word and also from his/her attitude towards his/ her utterance and/ or towards the interlocutors (Leech, 1990: 14). For example dwelling, house, home, abode, residence have the same denotation but different connotations.
Given their highly individual nature, connotations seem to be unrepeatable but, on the other hand, in many instances, the social nature of individual experience makes some connotative shades of meaning shared by practically all the speakers of a language. It is very difficult to draw a hard line between denotation and connotation in meaning analysis, due to the fact that elements of connotation are drawn into what is referred to as basic, denotative meaning. By taking into account connotative overtones of meaning, its analysis has been introduced a new dimension, the pragmatic one.
Talking about reference involves talking about nominals- names and noun phrases-. They are labels for people, places, etc. Context is important in the use of names; names are definite in that they carry the speaker’s assumption that his/ her audience can identify the referent (Saeed, 1997: 28).
One important approach in nominals’ analysis is the description theory (Russel, Frege, Searle). A name is taken as a label or shorthand for knowledge about the referent, or for one or more definite descriptions in the terminology of philosophers. In this theory, understanding a name and identifying the referent are both dependent on associating the name with the right description.
e. g. Christopher Marlowe / the writer of the play Dr. Faustus / the Elizabethan playwright murdered in a Deptford tavern.
Another interesting approach is the causal theory (Devitt, Sterelny, 1987) and based on the ideas of Kripke (1980) and Donnellan (1972). This theory is based on the idea that names are socially inherited or borrowed. There is a chain back to the original naming/ grounding. In some cases a name does not get attached to a single grounding. It may arise from a period of repeated uses. Sometimes there are competing names and one wins out. Mistakes can be made and subsequently fixed by public practice. This theory recognizes that speakers may use names with very little knowledge of the referent, so it stresses the role of social knowledge in the use of names. The treatment chosen for names can be extended to other nominals like natural kinds (e. g. giraffe, gold) that is nouns referring to classes which occur in nature.

B. Conceptual/ Representational Theory of Meaning
It proposes to define meaning in terms of the notion, the concept or the mental image of the object or situation in reality as reflected in man’s mind. Semantic studies, both traditional and modern, have used mainly such conceptual definitions of meaning, taking it for granted that for a correct understanding of meaning, it is necessary to relate it to that reflection in our minds of the general characteristics of objects and phenomena. Even Bloomfield refers to general characteristics of an object/ situation which is ‘linguistically relevant’.
On the other hand, complete identification of meaning with the concept or notion is not possible either. This would mean to ignore denotation and to deprive meaning of any objective foundation. More than that, languages provide whole categories of words-proper names, prepositions, conjunctions- for which no corresponding notions can be said to exist. Even in the case of notional words, the notion, the concept may be regarded as being both ‘wider’ and ‘narrower’ than meaning. A notion, concept has a universal character, while the meaning of a word is specific, defined only within a given language (Chiţoran, 1973: 32-33).
Signification and Sense. Meaning should be defined in terms of all the possible relations characteristic of language signs. The use of a linguistic sign to refer to some aspects of reality is a semiotic act. There are three elements involved in any semiotic act- the sign, the sense, the signification.
Two distinguishable aspects of the content side of the sign can be postulated- its signification, the real object or situation denoted by the sign, i. e. its denotation and a sense which expresses a certain informational content on the object or situation. The relation between a proper name and what it denotes is called name relation and the thing denoted is called denotation. ‘A name names its denotation and expresses its sense.’ (Alonso Church)
Extensional and Intensional Meaning. The definition of meaning by signification is called extension in symbolic logic (Carnap, 1960) and what has been called sense is equivalent to intension. Extension stands for the class of objects corresponding to a given predicate, while intension is based on the property assigned to the predicate (E. Vasiliu, 1970).
e. g. They want to buy a new car. (intensional meaning)
There is a car parked in front of your house. (extensional meaning)
C. The Trapezium of Heger.
Klaus Heger in his article Les bases metodologiques de l’onomasiologie proposes a trapezium- like variant, which allows him to introduce new distinctions. Heger noticed – as Greimas, adept of the triadic conception agreed- that signifiant + signifie i. e. concept is different from the linguistic sign, because the content of an expression is a semasiologic field, which is made up of more than one concept or mental object. In its turn a concept can be expressed by means of several signifiants.
The model of Heger gives him the possibility to analyse the content, making place for sememes and semes. Extralinguistic reality has two levels- the logical and/or psychological level and the level of the external world (C. Baylon, P. Fabre, 1978: 132).

The term moneme (A. Martinet) is also used by Heger and represents the minimal unit endowed with signification; a moneme is made up of morphemes which are in a limited number and it also represents a lexeme, the number of lexemes in a language being virtually infinite. In conclusion, a moneme is at the same time form of expression like phonemes and form of content like sememes. It is significant and signified. The signified depends on the structure of the language, but the concept on the right side of the trapezium is independent.

The onomasiology starts from the concept and tries to find the linguistic relations for one or several languages. It tries to find monemes which by means of their significations or sememes express a certain concept. An onomasiological field reprewsents the structure of all the sememes belonging to different signified, so to different monemes, but making up one concept.
Semasiology analyses a signified associated by co- substantiality to one moneme; so we deal with multiple significations or sememes.

Kurt Baldinger (1984: 131) comments on Heger’s trapezium, analysing the succesive stages from the substance of expression level to the final content level.
II. Dimensions of Meaning.
1. Dimensions of Meaning. Meaning is so complex and there are so many factors involved in it, that a complete definition would be impossible. We are dealing with a plurality of dimensions characteristic of the content side of linguistic signs (Chiţoran, 1973: 37).
There is a first of all a semantic dimension proper, which covers the denotatum of the sign including also information as to how the denotatum is actually referred to, from what point of view it is being considered. The first aspect is the signification, the latter is its sense.
e. g. Lord Byron/ Author of Child Harold have similar signification and different senses.
He is clever. /John is clever . He and John are synonymous expressions if the condition of co- referentiality is met.
The logical dimension of meaning covers the information conveyed by the linguistic expression on the denotatum, including a judgement of it.
The pragmatic dimension defines the purpose of the expression, why it is uttered by a speaker. The relation emphasized is between language users and language signs.
The structural dimension covers the structure of linguistic expressions, the complex network of relationships among its component elements as well as between it and other expressions.
2. Types of Meaning. Considering these dimensions, meaning can be analyzed from different perspectives, of which G. Leech distinguished seven main types (Leech, 1990: 9).
a. Logical/ conceptual meaning, also called denotative or cognitive meaning, is considered to be the central factor in linguistic communication. It has a complex and sophisticated organization compared to those specific to syntactic or phonological levels of language. The principles of contrastiveness and constituent structure – paradigmatic and syntagmatic axes of linguistic structure- manifest at this level i. e. conceptual meaning can be studied in terms of contrastive features.
b. Connotative meaning is the communicative value an expression has by virtue of what it refers to. To a large extent, the notion of reference overlaps with conceptual meaning. The contrastive features become attributes of the referent, including not only physical characteristics, but also psychological and social properties, typical rather than invariable. Connotations are apt to vary from age to age, from society to society.
e. g. woman [capable of speech] [experienced in cookery]
[frail] [prone to tears]
[non- trouser- wearing]
Connotative meaning is peripheral compared to conceptual meaning, because connotations are relatively unstable. They vary according to cultural, historical period, experience of the individual. Connotative meaning is indeterminate and open- ended that is any characteristic of the referent, identified subjectively or objectively may contribute to the connotative meaning.
c. In considering the pragmatic dimension of meaning, we can distinguish between social and affective meaning. Social meaning is that which a piece of language conveys about the social circumstances of its use. In part, we ‘decode’ the social meaning of a text through our recognition of different dimensions and levels of style.
One account (Crystal and Davy, Investigating English Style) has recognized several dimensions of socio-linguistic variation. There are variations according to:
– dialect i. e. the language of a geographical region or of a social class;
– time , for instance the language of the eighteenth century;
– province/domain I. e. the language of law, science, etc.;
– status i. e. polite/ colloquial language etc.;
– modality i. e. the language of memoranda, lectures, jokes, etc.;
– singurality, for instance the language of a writer.
It’s not surprising that we rarely find words which have both the same conceptual and stylistic meaning, and this led to declare that there are no ‘true synonyms’. But there is much convenience in restricting the term ‘synonymy’ to equivalence of conceptual meaning. For example, domicile is very formal, official, residence is formal, abode is poetic, home is the most general term. In terms of conceptual meaning, the following sentences are synonymous.
e. g. They chucked a stone at the cops, and then did a bunk with the loot.
After casting a stone at the police, they absconded with the money.
In a more local sense, social meaning can include what has been called The illocutionary force of an utterance, whether it is to be interpreted as a request, an assertion, an apology, a threat, etc.
d. The way language reflects the personal feelings of the speaker, his/ her attitude towards his/ her interlocutor or towards the topic of discussion, represents affective meaning. Scaling our remarks according to politeness, intonation and voice- timbre are essential factors in expressing affective meaning which is largely a parasitic category, because it relies on the mediation of conceptual, connotative or stylistic meanings. The exception is when we use interjections whose chief function is to express emotion.
e. Two other types of meaning involve an interconnection on the lexical level of language. Reflected meaning arises in cases of multiple conceptual meaning, when one sense of a word forms part of our response to another sense. On hearing, in a church service, the synonymous expressions the Comforter and the Holy Ghost, one may react according to the everyday non- religious meanings of comfort and ghost. One sense of a word ‘rubs off’ on another sense when it has a dominant suggestive power through frequency and familiarity. The case when reflected meaning intrudes through the sheer strength of emotive suggestion is illustrated by words which have a taboo meaning; this taboo contamination accounted in the past for the dying- out of the non- taboo sense; Bloomfield explains in this way the replacement of cock by rooster.
f. Collocative Meaning consists of the associations a word acquires on account of the meanings of words which tend to occur in its environment/ collocate with it.
e. g. pretty girl/ boy/ flower/ color
handsome boy/ man/ car/ vessel/ overcoat/ typewriter .
Collocative meaning remains an idiosyncratic property of individual words and it shouldn’t be invoked to explain all differences of potential co- occurrence. Affective and social meaning, reflected and collocative meaning have more in common with connotative meaning than with conceptual meaning; they all have the same open- ended, variable character and lend themselves to analysis in terms of scales and ranges. They can be all brought together under the heading of associative meaning. Associative meaning needs employing an elementary ‘associationist’ theory of mental connections based upon contiguities of experience in order to explain it. Whereas conceptual meaning requires the postulation of intricate mental structures specific to language and to humans, and is part of the ‘common system‘ of language shared by members of a speech community, associative meaning is less stable and varies with the individual’s experience. Because of so many imponderable factors involved in it, associative meaning can be studied systematically only by approximative statistical techniques. Osgood, Suci and Tannenbaum (The Measurement of Meaning, 1957), proposed a method for a partial analysis of associative meaning. They devised a technique – involving a statistical measurement device, – The Semantic Differential -, for plotting meaning in terms of a multidimensional semantic space, using as data speaker’s judgements recorded in terms of seven point scales.
Thematic Meaning means what is communicated by the way in which a speaker/ writer organizes the message in terms of ordering, focus or emphasis. Emphasis can be illustrated by word- order:
e.g. Bessie donated the first prize.
The first prize was donated by Bessie.
by grammatical constructions:
e. g. There’s a man waiting in the hall.
It’s Danish cheese that I like best.
by lexical means:
e. g. The shop belongs to him
He owns the shop.
by intonation:
e. g. He wants an electric razor.
Conclusions
a. meaning, as a property of linguistic signs, is essentially a relation- conventional, stable, and explicit- established between a sign and the object in referential definitions, or between the sign and the concept/ the mental image of the object in conceptual definitions of meaning;
b. an important aspect of meaning is derived from the use that the speakers make of it – pragmatic meaning, including the attitude that speakers adopt towards the signs;
c. part of the meaning of linguistic forms can be determined by the position they occupy in a system of equivalent linguistic forms, in the paradigmatic set to which they belong- differential/ connotative meaning;
d. equally, part of the meaning can be determined by the position a linguistic sign occupies along the syntagmatic axis- distributional/ collocative meaning;
e. meaning cannot be conceived as an indivisible entity; it is divisible into simpler constitutive elements, into semantic features, like the ones displayed on the expression level of language.
1. Conceptual Meaning Logical, cognitive or denotative content
Associative meaning 2. Connotative Meaning What is communicated by virtue of what language refers to
3. Social Meaning What is communicated of the social circumstances of language use
4. Affective Meaning What is communicated of the feelings and attitudes of the speaker/ writer
5. Reflected Meaning What is communicated through association with another sense of the same expression
6. Collocative Meaning What is communicated through association with words tending to occur in the environment of another word
7. Thematic Meaning What is communicated by the way in which the message is organized in terms of order and emphasis

Topics for discussion and exercises
1. Characterize the referential theories of meaning.
2. Define the terms referent, extension, denotation, connotation. Give examples to illustrate the definitions.
3. Identify and comment on the type of meaning of the bold words in terms of extension and intension
An Opera Theatre in her town is her dream.
They are signing the contract.
‘Have you met the Pope ‘ ‘I have never met Giovanni Paolo II’.
I wanted to find a nice pair of glasses but there wasn’t any cheap enough.
Since he saw that film, he’s always been afraid of ghosts.
Ann was sad. She didn’t answer my greeting.
He bought a bar of chocolate.
Zorro is his favorite hero.
They have no money to travel abroad.
Every year, the mayor delivers a speech in the town square.
What we need is a group of volunteers.
4. Give examples for each type of meaning in Leech’s classification.
Chapter III.
MOTIVATION OF MEANING

Ferdinand de Saussure’s apodictic statement: “the linguistic sign is arbitrary” in the sense that there is no direct relationship between the sound sequence (the signifiant) and the “idea” expressed by it (signifié) is taken for granted in the study of language. The resumption of the discussion on the arbitrary character of the linguistic sign in the late thirties and early forties proved however that the problem is not as simple as it might seem. There are numerous words in all languages in which a special correlation may be said to exist between meaning and sound. These words include in the first place interjections and onomatopoeia, which are somehow imitative of non-linguistic sounds as well as those instances in which it can be said that some sounds are somehow associated with certain meanings, in the sense that they suggest them. This latter aspect is known as phonetic symbolism.
But in addition to these cases which still remain marginal in the language, there is also another sense in which the meaning of words may be said to be related to its form, namely the possibility of analyzing linguistic signs by reference to the smaller meaningful elements of which they are made up. Indeed, derivative, complex and compound words are analyzable from the point of view of meaning in terms of their constituent morphemes.
It is obvious that while the general principle remains valid, namely that there is no inherent reason why a given concept should be paired to a given string of sounds, it is the linguist’s task to examine those instances, when it is possible to say something about the meaning of a linguistic sign by reference to its sounds and grammatical structure, in other words, it is necessary to assess the extent to which there is some motivation in the case of at least a number of words in the language.
Ullmann (1957) made a distinction between opaque and transparent words. In the latter case of transparent words, Ullmann discusses three types of motivation: phonetic, grammatical and semantic (motivation by meaning, as in the case of “breakfast”, whose meaning can be derived from the meaning of its component elements).
There are two main types of linguistic motivation already postulated by de Saussure: absolute and relative motivation.

1. Absolute motivation
Absolute motivation includes language signs whose sound structure reproduces certain features of their content. Given this quasi-physical resemblance between their signifiant and their signifié, these signs are of an iconic or indexic nature in the typology of semiotic signs, although symbolic elements are present as well in their organization:
There are several classes of linguistic signs, which can be said to be absolutely motivated:
(i) Interjections. It would be wrong to consider, as is sometimes done, that interjections somehow depict exactly the physiological and psychological states they express. The fact that interjections differ in sound from one language to another is the best proof of it. Compare Romanian au! aoleu! vai! etc. and English ouch!, which may be used in similar situations by speakers of the two languages.
(ii) Onomatopoeia. This is true of imitative or onomatopoeic words as well. Despite the relative similarity in the basic phonetic substance of words meant to imitate animal or other sounds and noises, their phonological structure follows the rules of pattern and arrangement characteristic of each separate language. There are instances in which the degree of conventionality is highly marked, as evidenced by the fact that while in English a dog goes bow-wow, in Romanian it goes ham-ham. Also, such forms as English whisper and Romanian şopti are considered to be motivated in the two languages, although they are quite different in form.
(iii) Phonetic symbolism. Phonetic symbolism is based on the assumption that certain sounds may be associated with particular ideas or meanings, because they somehow seem to share some attributes usually associated with the respective referents. The problem of phonetic symbolism has been amply debated in linguistics and psychology and numerous experiments have been made without arriving at very conclusive results.
It is quite easy to jump at sweeping generalizations starting from a few instances of sound symbolism.
Jespersen attached particular attention to the phonetic motivation of words and tried to give the character of law to certain sound and meaning concordances. He maintained for instance, on the basis of ample evidence provided by a great variety of languages, that the front, close vowel sound of the [i] type is suggestive of the idea of smallness, rapidity and weakness. A long list of English words: little, slim, kid, bit, flip, tip, twit, pinch, twinkle, click, etc. can be easily provided in support of the assumption, and it can also be reinforced by examples of words from other languages: Fr. petit, It. piccolo, Rom. mic, etc. Of course, one can equally easily find counter examples – the most obvious being the word big in English – but on the whole it does not seem unreasonable to argue that a given sound, or sequence of sounds is associated to a given meaning impression, although it remains a very vague one.
Sapir (1929) maintained that a contrast can be established between [i] and [a] in point of the size of the referents in the names of which they appear, so that words containing [a] usually have referents of larger size. Similar systematic relations were established for consonants as well.
Initial consonant clusters of the /sn/, /sl/, /fl/ type are said to be highly suggestive of quite distinctive meanings, as indicated by long lists of words beginning with these sounds.
2. Relative motivation
Relative motivation. In the case of relatively motivated language signs, it is not the sounds which somehow evoke the meaning; whatever can be guessed about the meaning of such words is a result of the analysis of the smaller linguistic signs which are included in them. Relative motivation involves a much larger number of words in the language than absolute motivation. There are three types of relative motivation: motivation by derivation; by composition and semantic motivation.
An analysis of the use of derivational means to create new words in the language will reveal its importance for the vocabulary of a language. The prefix {-in}, realized phonologically in various ways and meaning either (a) not and (b) in, into, appears in at least 2,000 English words: inside, irregular, impossible, incorrect, inactive etc.
Similarly, the Latin capere (“take”) appears in a great number of English words: capture, captivity, capable, reception, except, principal, participant, etc.
It is no wonder that Brown (1964) found it possible to give keys to the meanings of over 14,000 words, which can be analyzed in terms of combinations between 20 prefixes and 14 roots. Some of his examples are given below:
Words Prefix Common Meaning Root Common Meaning
1. Precept pre- before capere take, seize
2. Detain de- away, down tenere hold, have
3. Intermittent inter- between, among mittere send
4. Offer ob- against ferre bear, carry
5. Insist in- into stare stand
6. Monograph mono- alone, one, graphein write
7. Epilogue epi- upon legein say, study of
8. Aspect ad- to, towards specere see
This table alone is sufficient to indicate the importance of relative motivation for the analysis of meaning.
It is obvious that the lexicon of a language presents items which differ in the degree to which their meaning can be said to be motivated; while some are opaque (their sound give no indication of their meaning), others are more or less transparent, in the sense that one can arrive at some idea of their meaning by recourse to their phonetic shape or to their derivational structure or to some semantic relations which can be established with other words in the language.
In Précis de sémantique française (1952), Ullman suggested several criteria of semantic structure which enabled him to characterize English as a “lexical language”, as opposed to French which is a more “grammatical” one: the number of arbitrary and motivated words in the vocabulary; the number of particular and generic terms; the use of special devices to heighten the emotive impact of words. Three other criteria are based on multiple meaning (patterns of synonymy, the relative frequency of polysemy, and the incidence of homonymy) and a final one evaluates the extent to which words depend on context for the clarification of their meaning. This is an area of study which could be continued with profitable results for other languages as well.

Bibliography:
Chiţoran, Dumitru. 1973. Elements of English Structural Semantics, Bucureşti, Ed. Didactică şi Pedagogică.
Exercises:
1. Give examples of words which are absolutely motivated.
2. Analyse the following words in terms of relative motivation: rowboat, impermeability, wholesaler, pan-African, childless, playing-field, incredible, scare-crow, counter-attack, imperfect, overdose, shareholder, caretaker, salesman, foresee, misunderstanding.
3. Give examples of words build with the help of the following prefixes: bi-, in-, mis-, de-, anti-, non-, out-, super-, dis-, mal-, a-, en-, over-.
4. Analyze the following blends in point of their relative motivation: sportcast, smog, telescreen, mailomat, dictaphon, motel, paratroops, cablegram, guestar, transistor.
5. Write the word forms of the following words and analyze them in terms of relative motivation: move, comment, place. Consider Saussure’s types of associations and find possible associations among the word forms that you previously found.
Chapter IV
STRUCTURAL APPROACHES TO THE STUDY OF MEANING
1. COMPONENTIAL ANALYSIS

Though structuralism in linguistics should be connected to structuralism in other sciences, notably in anthropology, it should also be regarded as a result of its own inner laws of development as a science.
Generally, structuralist linguistics may be characterised by a neglect of meaning, but this must not lead to the conclusion that this direction in linguistics has left the study of meaning completely unaffected. Structural research in semantics has tried to answer two basic guestions:
a) – is there a semantic structure/system of language, similar to the systemic organisation of language uncovered at other levels of linguistic analysis (phonology and grammar) ?
b) can the same structure methods which have been used in the analysis of phonological and grammatical aspects of languages be applied to the analysis of meaning ?
In relation to question a), the existence of some kind of systemic organisation within the lexicon of a language is taken for granted. F de Sanssure pointed aut that the vocabulary of a language cannot be regarded as a mere catalogue. But this aaceptance does not mean it is an easy job to prove the systematic character of the lexicon. First of all, it would mean the study of the entire civilization it reflects and secondly, given the fluid and vague nature of meaning, semantic reality must be analysed without recourse to directly observable entities as it happens in case of sound and grammatical meaning.
One solution was to group together those elements of the lexicon which form more or less natural series. Such series are usually represented by kinship terms, parts of the human body, the term of temporal and spatial orientation,etc, that can be said to reveal a structural organisation. Structural considerations were applied to terms denoting sensorial perceptions: colour, sound , swell, taste, as well as to terms of social and personal appreciation.
The existence of such semantic series, the organisation of words into semnatic fields justified the structural approach to the study of lexicon.
Hjelmslev conditioned the existencee of system in language by the existence of paradignes so that a structural description is only possible where paradigmes are revealed.But the vocabulary , as an open system, with a variable number of elements, does not fit such a description unless the definition of system broadens. Melcuk (1961) stated that a set of structurally organised objects forms a system if the objects can be described by certain rules, on condition that the number of rules is smaller than the number of objects. Constant reference to phonology, in terms of distinguishing between relevant and irrelevant in the study of meaning has led to applying methods pertaining to the expression level of language to its content level as well.
Some linguistic theories, mainly the Gloosemantic School, take it for granted that there is an underlying isomorphism between the expression and content levels of language. Accordingly they consider it axiomatic to apply a unique method of analysis to both levels of language. Hjelmslev distinguishes between signification and sense and deepens this distinction on the basis of a new dichotomy postulated by glossematics : form and substance. While the sense refers to the substance of content, signification refers to its form or structure. The distinction signification/sense can be analysed in term of another structuralist dichotomy: invariant/variant. Significations represent invariant units of meaning while the sense are its variants. There is a commutation relation between significations as invariants, and a substitution one between senses as variants. An example is given below :
Romanian English Russian
palma
mana
hand
pyka
brat arm
Since significations as invariants find their material manifestation in senses as their invariants, in terms of glossematics, a theory of signification stands for content form alone, so signification is no more semantic than other aspects of content form dealt with by grammar. It follows that only a theory of the sense (substance of content) could be the object of study of semantics(Chitoran, 1973:48).
In Hjelmslev’s opinion, sense is characteristic of speech, not of language, pertains to an empirical level, so below any interest of linguistics. Any attempt to uncover structure or system at the sense level can be based on the collective evaluation of sense. For Hjelmslev, lexicology is a sociological discipline which makes use of linguistic material : words. This extreme position is in keeping with the neopositivist stand adopted by glossematics, according to which form has primacy over substance, that language is form, not substance and what matters in the study of meaning is the complex network of relations obtaining among linguistic elements.
Keeping in mind the basic isomorphism between expression and content, it is essential to emphasize some important differences between the two language levels:
– the expression level of language implies sequentiality, a development in time (spoken language) or space (written language); its content level is characterised by simultaneity;
– the number of units to be uncovered at the expression level is relatively small, and infinitely greater at the content level.
It is generaly accepted that the meanings of a word are also structured, that they form microsystems, as apposed to the entire vocabulary which represents the lexical macrosystem. The meanings of a lexical element display three levels of structure, starting from a basic significative nucleus, a semantic constant (Coteanu, 1960) which represents the highest level of abstraction in the structuration of meaning. Around it different meanings can be grouped (the 2 nd level). (Chiţoran, 1973:51)
The actual uses of a lexical item, resulting from the individualising function of words (Coteanu, 1960) belong to speech. Monolingual dictionaries give the meanings of a lexical item abstracted on the basis of a wide collection of data. As far as the semantic constant is concerned, its identification is the task of semnatics and one way of doing that is by means of the Componential Analysis.
Componential Analysis assumes that all meanings can be further analysed into distinctive semantic features called semes, semantic components or semantic primitives, as the ultimate components of meaning. The search for distinctive semantic features was first limited to lexical items which were intuitively felt to form natural structures of a more ar less closed nature. The set kinship terms was among the first lexical subsystems to be submitted to componential analysis :
father [+male][+direct line] [+older generation]
mother [-male][+direct line] [+older generation]
son [+male][+direct line] [-older generation]
daughter [_male][+direct line] [-older generation]
uncle[+male][_direct line] [+older generation]
aunt [-male] [-direct line] [+older generation]
nephew [+male] [-direct line] [-older generation]
niece [-male] [-direct line] [-older generation]
It is evident than there exist the same hierarchy of units and the same principle of structuring lower level units into higher level ones (Pottier, 1963):
Expression Content
Distinctive feature pheme (f) seme (s)
Set of distinctive features phememe(F)
(a set of pheme) sememe (S)
(A set of semes)
The formalization of a set of
Distinctive features phoneme(P)
(the formalization of a phememe) lexeme(L)
formalization of a sememe
The sememes are arrived at by comparing various lexical items in the language. Starting from the dictionary definitions, the semantic features encountered in case of furniture intended for siting are :
Semantic feature/
Lexical item for sitting with back with support
for arms for more
people
upholstered
Stool + – – – *
Chair + + – – *
Armchair + + + – *
Bench + + * + –
Sofa + + * + +
*the given feature(present/absent)is not relevant .
On the content level an archilexeme will result from the neutralization of a lexemic opposition. In this case the more general term chair can be the archilexeme, or another lexical item can be chosen-seat.
Glossematies represents the point of departure for an American linguistic theory, the statificational theory of language (Sidney Lamb, 1964,1966). He included a semantic theory in his general linguistic theory. This semantic component has the form of a separate level of language (stratum) the sememic one. Lamb’s semantic theory is based on the assumption that there is a structuralization of meaning characteristic of all languages. While before him words were related directly to their denotata or significata. Lamb suggests the insertion of a new statum ‘sememics’, between language and the outside world in order to delimit what is linguistically relevant on the content level from what is not. The sememic statum is inserted between the lexemic (lower) and the semantic (higher) strata.Its elementary unit is the semon(=the minimal unit of the semantic stratum such that its components are not representations of the components of the semantic statum sememes may be accounted for by general construction rules, the combination of semons must be listed individually for each sememe. Evidence is formed both for diversification (semo-lexemic) and neutralization (lexosememic) between the two strata.
Sememic stratum S of colours; giving out/reflecting much light

Lexemic stratum L bright L vivid L intense

Sememic stratum S quick-witted, clever

Lexemic stratum L bright Lgifted L clever L capable
Sememeic statum S/piece of wood s/on the ship s/group of people s/(food)

Lexemic stratum board
The first is accounting for the semasiological direction, the second for the onomasiological direction (from denotata and significata to a linguistic form-explaining synonymy). In the process of neutalization which accounts for polysemy, one lexema is connected to several sememes in an either-or type of relationship. But the lexeme/lamb/is connected both to the sememe/sheep/ and the sememe /young/. A given lexeme may connect first to several units in an either-or relationship, which in turn may connect to several sememes in a both-and relationship. The intermediate units between the lexeme and the sememes are called by Lamb sememic signs.
/male/
Sememic stratum /unmarried /owner of the / ± male /
person / 1 st Acad. Degree/
(intermediate) /unmarried /university /young
sememic sign man/ graduate/ knight/
(sememe)
lexemic stratum bachelor
By expressing the meanings of individual items in terms of combinations of features, we obtain the componential definitions of the items concerned. They can be regarded as formalized dictionary definitions :
man + HUMAN + ADULT + MALE
The dimensions of meaning will be termed semantic oppositions. The features of opposition are mutually defining.
+ (marked)
– (negative, unmarked)
Not all semantic contrasts are binary In fact componential analysis assumes that meanings are organised in multi-dimensional contrasts. Taxonomic (hierarchical arrangement of categories) oppositions can be :
– binary : dead # alive
– multiple : gold # copper # iron # mercury etc.
The link between componential analysis and and basic statements is made through the mediation of hyponymy (inclusion) and incompatibility. So basic logical relationships (entailment, inconsistency) can be defined in terms of hyponymy and incompatibility (Leech, 1990:97):
e.g. The secretary is a woman entails The secretary is an adult.
I meet two boys entails I met two children.
Justifying componential analysis by following out its logical consequences in terms of basic statements implies giving a certain priority to sentence meaning over word-meaning, so truth-falsehood properties of sentence meanings are the surest basis for testing a description of meaning: scared and frightened would be considered as synonyms in terms of their truth value and would be perceived as differing in terms of stylistic meaning +/- colloquial.
The features of different semantic oppositions can be combined. Is it true that every dimension is variable completely independent of all the other ?
/+ human/ /+ adult/ /+male/ are independently variable
/+animate/ combines with /+countable/

/+ animate/ combines with /± male/ but [+male] implies [+animate]
Redundancy rules add features which are predictable from the presence of other features and are therefore in a sense redundant to an economical semantic interpretation. Such rules are found in phonology and syntax. Indirect relation of incompatibility and hyponymy can be established through redundancy rules: man and book are incompatible in meaning.
Hence, X is a man and X is a book are inconsistent statements. Redundancy rules are important for extending the power of componential analysis to account for basic statements. Certain features and oppositions can be regarded as more important than others in the total organisation of the language. The oppositions ± concrete and + countable have many other oppositions dependent on them and so they are in key positions as it happens with the feature + animate. (G. Leech, 1990:111).
Binary oppositions frequently have marked and unmarked terms. That is, the terms are not entirely of equivalent weight, but one (the unmarked) is neutral or positive in contrast to the other.
e.g. book books
petit petite
duck drake
long short
Markedness is definable as a relation between form and meaning : if two words contrast on a single dimension of meaning, the unmarked one is the one which can also apply neutrally to the whole dimension. A positive-negative bias is inherent to the semantic opposition. Often the marked term is indicated by a negative suffix or prefix : happy-unhappy, useful-useless. People tend to respond more quickly to unmarked than to marked terms. This could be explained by their tendency to look on the bride side of life and associate unmarkedness with ‘good’ evaluations and markedness with ‘bad’ ones (Leech, 1990:114).
There is also a factor of bias in relative oppositions but this could be explained in terms of dominance rather than markedness. We prefer to use the dominant term before the other or to use it alone.
parent/child see –
own/belong to hit –
in front/behind have –
Markedness and dominance vary in strength (they can grow weak even become inexistent left/right) and are also subject to contextual influences.

Criticisms of Componential Analysis. Componential analysis is considered by some linguists as a useful and revealing technique for demonstrating relation of meaning between words. At the same time, this theory of word-meaning has been criticised and G.Leech has tried to comment on the main criticisms :
1. It is said that componential analysis (CA) accounts for only someparts of a language’s vocabulary (those parts which are neatly organized). Componential analysis can be fitted into a more powerful model of meaning, with additional levels of analysis apart from CA. Semantic features need not be atomic contrastive elements, but may have an internal structure of their own, that is, the semantic features can be derived from configurations of other features. This recursive power of feature-creation is particularly important in considering metaphor. So, there is no need to postulate an indefinite proliferation of semantic oppositions.
2. It is often objected than CA suffers from a ‘vicious circle’ in that it merely explains one set of symbols (e.g. English words) by another set of symbols (which also turned out to be English words). The notation of symbols is arbitrary and the explanatory function of features is solely their role in the prediction of basic statements.
3. Another objection is that CA postulates abstract semantic entities (semantic features) unnecessarily. But the notation of CA is language-neutral, and so the same features, oppositions redundancy rules may explain meaning relation in many different languages.
4. Connected to that, it has been postulated that CA implies universal features of meaning and therefore relies on the strong assumption that the same semantic features are found in all languages. CA fits in well with a ‘weak universalist’ position whereby semantic oppositions are regarded as language-neutral i.e. as conceptual contrasts not necessarily tied to the description of particular languages. Semantic analyses may be generalized from one language to another, but only to the extent that this is justified by translation equivalence.
5. It has also been claimed that CA is unexplanatory in that it does not provide for the interpretation of semantic features in terms of the real-world properties and objects that they refer to. For example + ADULT remains an abstract uninterpreted symbol unless we can actually specify what adults are like i.e. how decide when the feature + ADULT refers to something. To expect CA to provide an interpretation in this sense is to expect it to provide a theory not only of meaning, but of reference, or not only of conceptual meaning, but also of connotative meaning. CA cannot have this wider goal : it is meant to explain word sense, not the encyclopedic knowledge which must enter into a theory of reference.
6. The view that word-meanings are essentially vague, that determinate criteria for the reference of words cannot be given has received prominent support in philosophy and linguisties. Wittgenstein exemplified this with the word game : he could find no essential defining features of what constitutes a game and concluded that we know the meaning by virtue of recognizing certain ‘family resemblances between the activities it refers to. A more recent critique of the deterministic view of meaning is given by Labov (1973) who conducted an experiment in which subjects were invited to label pictures of more-or-less cup-like objects. There was a core of agreement as to what constituted a cup but there was also a peripheral gradient of disagreement and uncertainty. The conclusion is that cup, mug, bowl and similar words are defined in terms of ‘fuzzy sets of attributes’, that is sets of attributes of varying importance, rather than in terms of a clear-cut, unvarying set of features. We match candidates for ‘cuphood’ against a prototype or standard notion of cup. The vagueness is referential and does not affect componential analysis because it has to do with category recognition: the mental encyclopedia rather than the mental dictionary.
Another kind of variability of reference is presented by Lyons in case of three words: boy, girl, child in terms of a common feature – ADULT. This feature will require different interpretations in the three cases. Within the-ADULT category there is a further binary taxonomy distinguishing child from adolescent. –ADULT stands as a common factor in the meanings of boy, girl, child, puppy etc. but its referential interpretation is variable for reasons which are explicable in terms of the prototypic view of categories.
There have emerged three different levels at which word-meaning can be analysed.
– the word-sense as an entirety may be seen as a conceptual unit in its own right prepackaged experience (Leech, 1990:121);
– this unit may be subdivided into components/features by CA;
– both word-senses and features, representing prototypic categories can be broken down into fuzzy sets of attributes.

2. PARADIGMS IN LEXIC
The Semantic Field Theory
The idea of the organization of the entire lexicon of a language into a unitary system was for the first time formulated by Jost Trier. Actually, Trier continued two lines of thought. On the one hand, he was directly influenced by W. von Humboldt and his ideas of linguistic relativism. Wilhelm von Humboldt, influenced by the romanticism of the early 19th century advanced the theory that languages are unique, in that each language expresses the spirit of a people, its Volksgeist. Each language categorizes reality in different ways so that it may either help or hinder its speakers in making certain observations or in perceiving certain relations. Given the principle of relativism, it follows that the vocabularies of any two languages are anisomorphic, that there are no absolute one to one correspondences between two equivalent words belonging to two different languages. Humboldt made, also, the distinction between language viewed statically as an ergon and language viewed dynamically, creatively, as an energeia. Trier’s semantic fields are, accordingly, closely, integrated lexical systems in a dynamic state of continuous evolution.
The other line of thought which Trier continues springs from Ferdinand de Saussure’s structuralism, more specifically from the distinctions made by the latter between the signification, and value of lexical items. According to de Saussure, words have signification, in that they do mean something, positively, but they also have value, which is defined negatively by reference to what the respective words do not mean. Linguistic value is the result of the structural relationships of a term in the system to which it belongs. Thus, Trier postulated that no item in the vocabulary can be analyzed semantically unless one takes into account the bundle of relationships and oppositions it enters with the other words in a given subsystem or system. One cannot assess the correct meaning of “green” for instance, unless one knows the meaning of “red” and all the other colours in the system.
Colour terms are actually often used to illustrate the semantic field theory. Let us suppose that the field of colours, which physicists assure us forms a continuum, is covered by the following number of terms in two languages L1 and L2:
L1: x y z
L2: a b c d e
It is evident that no single term in any of the two languages covers exactly the same area of the spectrum; only “z” in L1 can be said to incorporate the whole of “e” in L2 although it covers a small part of the area covered by “d” as well.
English and Shona, a language spoken in Rhodesia, exhibit precisely the type of structural segmentation of the colour spectrum postulated above. While English have seven basic terms for colour (the first level of the hierarchy), red, orange, yellow, green, blue and purple, Shona has only three which are distributed roughly as follows: a first term “covers the range of English orange, red and purple, and a small part of blue; another term covers the area of green and most of blue” (Lamb 1969: 46). It is evident that the terms for colour are not equivalent in the two languages.
Evidently the linguistic field of colour terms is a favourable one for such an analysis. There is first of all a “metalanguage” provided by the science of physics to which one can report the words for colour. Secondly, the number of words, is quite limited and thus reductible to a restricted set of relationships.
But even in the case of the most elementary vocabulary one encounters a similar lack of correspondence. English sheep and French mouton are not the same since English makes use of another term mutton, to cover the entire area of meanings and uses covered by French mouton.
Trier advanced the idea, that vocabulary as a whole forms an integrated system of lexemes interrelated in sense, a huge mosaic with no loopholes or superposed terms since our concepts themselves cover the entire Universe. According to his dynamic conception of language viewed as “energeia”, Trier pointed out that the slightest change in the meaning of a term within a semantic field brings about changes in the neighbouring terms as well.
Any broadening in the sense of one lexeme involves a corresponding narrowing in the sense of one or more of its neighbours. According to Trier, it is one of the major failings of traditional diachronic semantics that it sets out to catalogue the history of changes in the meanings of individual lexemes atomistically, or one by one, instead of investigating changes in the whole structure of the vocabulary as it has developed through time. (Lyons 1977: 252).
The procedure followed by Trier in diachronic semantics is not one of comparing successive states of the total vocabulary (which would be hardly practicable). What he does is to compare the structure of a lexical field at time t1 with the structure of a lexical field at time t2.
Semantic fields with a more restricted number of terms are incorporated into larger ones, the latter are themselves structurated into even larger ones, until the entire lexicon of a language is integrated into a unitary system. In Trier’s opinion therefore semantic fields act as intermediaries between individual lexical entries, as they appear in a dictionary, and the vocabulary as a whole.
Despite their revolutionary character, Trier’s ideas on semantics found few followers and were consequently slow in being pursued and developed. This is normal in view of the important objections which can be raised to his theory.
One of the objections came from those who were reluctant to admit such a perfect organization of vocabulary into an interdependent and perfectly integrated system of elements which delimit each other like pieces in a jig-saw puzzle. Secondly, the linguistic relativism of Trier’s ideas, his contention about the influence of language upon thought was rightly considered as an instance of linguistic solipsism.
Much of the criticism leveled at semantic field theory originated from less philosophical considerations. It is quite difficult to outline the actual limits of a field, its “constant”, which subsequently enables one to analyze the terms incorporated in it. Also, the semantic field theory, if valid, accounts for only one type of relations contracted by lexical items – the paradigmatic ones, or, a full semantic description should include syntagmatic relations as well. In addition Trier’s theory does not seem to be related to any given grammatical theory.
Nevertheless, there were numerous attempts at developing the semantic field theory, most of them departing to a lesser or greater extent from Trier’s original ideas. L. Weisgerber for instance, continued the analysis of the semantic field of knowledge and understanding in Modern German while trying to incorporate the notion of semantic fields in his general theory of language (1953).
P. Guiraud (1956, 1962) developed the theory of the morpho-semantic field. The morpho-semantic field includes all the sound and sense associations radiating from a word; its homonyms and synonyms, all other words to which it may be related formally or logically, metaphorically, etc., as well as casual or more stable associations which can be established between objects designated by these words.
Walter von Wartburg and R. Hallig (1952) undertook a more ambitious task. They suggested a method of analysis based on the system of concepts which was meant to cover the entire vocabulary of a language and, since the general classification of concepts was supposed to have a general character, the vocabulary of any language could be incorporated into such a conceptual dictionary.
The method is entirely reminiscent of Roget’s Thesaurus in that it identifies lexical systems with logical systems of concepts. The outline of the system of concepts has three main components: A: The Universe; B: Man; and C: Man and the Universe. Each main component includes several classes of concepts (and accordingly, of words designating these concepts). Thus, component A includes the following four classes: I The sky and atmosphere; II. The Earth; III. The Plants; IV. The Animals.
Semantic fields are structural organizations of lexis which reflect a structuration of the content level of language. Hjelmslev and E. Coseriu (1968) considered that any semantic theory is valid only to the extent to which it arrives at paradigms on the content level of language.
Coseriu defined the semantic field as a primary paradigmatic structure of the lexic, a paradigm consisting in lexical units of content (lexemes), which share a continuous common zone of signification, being in an immediate opposition one to another. (Iliescu, Wald 1981: 39)
A semantic field should be understood in Trier’s original sense, namely as a zone of signification covered by a number of closely interrelated lexical items. In this respect the componential analysis of meaning (Goodenough, 1956) seems to be nearer the true concept of the semantic field.
Three main objections can be and have been raised with regard to the present state of the semantic field theory.
(a) Is it possible to analyze the entire vocabulary into semantically structured fields, or are they limited to certain parts of it only, namely to lexical items designating aspects of reality (especially man-made reality, the reality of artifacts) which by their own nature possess a certain structural organization?
(b) Closely related to objection (a) one can doubt the linguistic nature of semantic fields. Do they correspond to an internal organization of the vocabulary or are they organizations external to language?
(c) How can semantic fields be delimited? Is there an objective method of evaluating the range of a given field and the number of elements it includes?

Componential Analysis Applied in the Analysis of Semantic Fields
One of the most important tenets of modern semantics claims that the meanings of lexical items do not represent ultimate, indivisible entities; they are, on the contrary, analyzable into further components. This led to a method of approach in semantic analysis, appropriately called componential analysis, previously discussed in this chapter.
Componential analysis originally started as a method of analysing units belonging to a certain semantic field. The method was fruitfully applied in the study of kinship terms, colour terminology, military ranks and other fairly restricted domains of meaning.
Assuming that the meaning of a word is not an undivided entity, componential analysis provides for the decomposition of meanings into smaller significant features. Modeled on the analysis of phonemes into distinctive features, componential analysis is founded on the notion of semantic contrast: the units of a field are assumed to contrast simultaneously on different dimensions of meaning. The meanings of the field units complement each other constituing a paradigm. A paradigm will be defined as a set of linguistic forms wherein:
a) the meaning of every form has, at least one feature in common with the meaning of all other forms in the set;
b) the meaning of every form differs from that of every other form of the set, by one or more additional features.
The common feature of meaning of the set is called the root meaning. It defines the semantic area which is analyzed by the units of the field. The words in the field will be arranged into contrastive sets along different dimensions of meaning. Thus, just as /t/ and /d/ complement each other with respect to the dimension of voicing, old and young complement each other with respect to the conceptual dimension of age.
A dimension is an opposition of mutually exclusive features. The features of the dimension sex, presumably relevant in an analysis of kinship terms, are [+Male] and [+Female].
Any term of the paradigm will be defined componentially in terms of its coordinates in the paradigm. The componential definition of a word is a combination of features for several (or for all) dimensions of the paradigm.
In the componential definition of the meaning of a lexical item the linguist proceeds from extensional definition to intensional definitions. That is, starting his analysis of say, kinship terms, the linguist has to draw up the list of all the terms with kinship designation and, than, to specify for each of them the set of possible denotata (the set of contextual meanings or all the allosemes of the word).
The componential definition of a term may be taken to be an expression of its significatum. A componential definition is therefore an intensional definition, which specifies the distinctive features shared in common by all denotata designated by a given term.
It is a unitary, conjunctive definition implying that all the features are simultaneously present in every occurrence of the word.

Bibliography:
1. Chiţoran, Dumitru. 1973. Elements of English Structural Semantics, Buc.: Editura Didactică şi Pedagogică.
2. Iliescu, M. Wald, L. 1981. Lingvistica modernă în texte. Buc.: Reprografia Universităţii din Bucureşti.
3. Lyons, J. 1977. Semantics vol. I, Cambridge: Cambridge University Press.
4. Hulban, N. Luca-Lăcătuşu, T. Creţescu Kogălniceanu, C. 1983. Competenţă şi performanţă. Exerciţii şi teste de limbă engleză. Bucureşti: Ed. Ştiinţifică şi Enciclopedică.

TOPICS FOR DISCUSSION AND EXERCISES
1. For each of the following words try to establish sets of attributes that would distinguish it from its companions in the group :
cake, biscuit, bread, role, bun, cracker,
boil, fry, broil, sauté, simmer, grill, roast.
2. For each group of words given below, state what semantic property /-ies distinguish between the classes of a) and b) words. Do a)words and b)words share any semantic property ?
Example: a) widow, mother, sister, aunt, maid
b) widower, father, brother, uncle, valet
a) and b) are human
a) words are female and b) male

A) a) bachelor, man, son, pope, chief
b) bull, rooster, drake, ram
B) a) table, stone, pencil, cup, house, ship, car
c) milk, alchohol, rice, soup, mud
C) a) book, temple, mountain, road, tractor
b) idea, love, charity, sincerity, bravery, fear
D) a) walk, run, skip, jump, hop, swim
b) fly, skate, ski, ride, cycle, canoe
E) a) alleged, counterfeit, false, putative, accused
b) red, large, cheerful, pretty, stupid
3. Define the terms seme, sememe, lexeme. Give examples.
4. What is a semantic field?

Chapter V
LANGUAGE AS A CONCEPTUAL SYSTEM

Language is not only an instrument of communication. It is far more than this – it is the means by which we interpret our environment, by which we classify or “conceptualize” our experiences, by which we are able to impose structure on reality, so as to use what we have observed for present and future learning and understanding. Leech considers language, in its semantic aspect, as a conceptual system. Not as a closed, rigid, conceptual system which tyrannizes over the thought processes of its users, but as an open-ended conceptual system, one which “leaks”, in the sense that it allows us to transcend its limitations by various types of semantic creativity.
The first question which arises in whether language is a single conceptual system, or whether there are as many conceptual systems as there are languages. Although much of present-day thinking has tended to hypothesize a universal conceptual framework which is common to all human language, common observation shows that languages differ in the way they classify experience. A classic instance of this is the semantics of colour words. English (according to Berlin and Kay, Basic Color Terms, 1969) has a range of eleven primary colour terms (black, white, red, green, yellow, blue, brown, purple, pink, orange and grey), whereas the Philipine language of Hanunóo (according to Conklein, Hanunóo Colour Categories, 1955) makes do with four.
Conceptual boundaries often vary from language to language. Languages have a tendency to impose structure upon the real world by treating some distinctions as crucial, and ignoring others. The way a language classifies things is sometimes blatantly man-centred.

Linguistic Relativism and Semantic Universals
Semantic relativism and semantic universals are two conflicting points of view in relation to meaning. Both theses concern the relation between the structure of language and the structure of the universe. They represent in fact two different ways of interpreting the relation between the universe, as experienced by man, and language as a tool of expressing that experience. Ever since ancient times it has been maintained that the structure of language reflects more or less directly the structure of the Universe as well as the universal structure of the human mind (Mounin, 1963: 41). This was taken to be a precondition of interlingual communication as well as of the act of translation.
In terms of Hjelmslevian distinction between substance and form of the content, it was agreed that there may be different ways of segmenting substance, and an even richer variety in its form but the content itself, the world of experience remains basically the same.
Linguistic relativism. The axiomatic character of the statement which relates the structure of language to the structure of the universe as reflected in man’s mind, ceases to be commonly agreed upon when one begins to consider the nature of this relationship.
Wilhelm von Humboldt in the first half of the 19th century, and many philosophers and linguists after him, assigned language a much more active role, regarding it not as a passive carrier of thought, but, in a very direct way as a moulder of it. In their opinion, language imposes upon thought its own system of distinctions, its own analysis of objective reality. These ideas remained unheeded by linguists until the advent of European structuralism. The key idea in Saussurean linguistics namely that language signs have no meaning or “value” outside the system to which they belonged, fits perfectly the principle of linguistic relativism. Trier and particularly Hjelmslev consider that each language structures reality in its own way and by doing so, creates an image of reality which is not a direct copy of it. Language is the result of the imposition of same form upon an underlying substance.
Quite independently, and emerging mainly from current observation in linguistic anthropological research on Amerindian languages, conducted by Fr. Boas, similar ideas were expressed by E. Sapir and B. L. Worf in America. Linguistic determinism has come to be often referred to as the Sapir-Worf hypothesis. For Sapir (1921) and Worf (1956) objective reality is an undifferentiated continuum which is segmented by each language in a different way. We obtain a vision of nature, of reality which is by and large pre-determined by our mother tongue. Each language is a vast system of structures, different from that of others in which are ordered culturally all forms and categories by means of which the individual not only communicates but also analyzes nature, grasps or neglects a given phenomenon or relation, in means of which he molds his manner of thinking and by means of which he builds up the entire edifice of his knowledge of the world. Worf provided ample evidence from Amerindian languages of how languages segment reality differently by neglecting aspects which are emphasized in other languages. In Europe linguists as Benveniste (1958) and Martinet, in analyzing the relationship between categories of thought and categories of language, are unanimous not only in pointing out a basic parallelism between the two, but also in assigning to linguistic categories a primary role. The linguistic structure conditions, albeit in an unconscious way, man’s knowledge of the world, his spiritual and philosophical experience.
Linguistic relativism or determinism in its extreme variant, which maintains that people’s knowledge of the world, the categorization of external experience is totally determined by the structure of language which imposes its particular form upon it, has been criticized. Various arguments can be advanced against the Sapir-Worf position. The idea that language systems have no points in common at all, and are completely untranslatable is refuted by empirical evidence. The fact that speakers of a given language are able to learn the vocabularies of other language, and, indeed, other languages as a whole, is the best proof of it. Also, a single language often has alternative conceptualizations of the same phenomenon: in English, for instance, human beings can be categorized by age into “children”, “adolescents” and “adults” or alternatively, into “majors” and “minors”. Furthermore, if we draw a distinction between meaning and reference, we can say that even though there is no corresponding concept in one’s own language for a concept in another language, one can nevertheless provide a description of its referent (Leech 1990: 27). The differences in environment, climate, cultural development, etc., among various linguistic communities may be very great, but basically, human societies are linked by a common biological history. The objective reality in which they live is definitely not identical but it is by and large similar. Man’s universe is basically a Universe made up of things and he is constantly confronted with them, obliged to communicate about them, to define himself in relation to them. This is basic to all human societies. Various language systems are not therefore untranslatable.
The problem of translatability or rather degrees of translatability may be discussed appropriately with reference to the notion of cultural overlap. Cultures are not linguistically bound; in other words, languages and cultures are not co-terminous. Linguistic boundaries do not coincide with cultural ones. There is always a certain degree of cultural overlap between two language communities.
On the whole, similarities among languages are more important and more numerous than the differences among them. These differences can be explained in terms of cultural differences between the respective language communities.
Second language learning too seems to support this point of view. Words denoting objects, structures and features situated in an area of cultural overlap are among the first to be learned, and with no apparent difficulty. Their acquisition seems to form the foundation on which the other words in the new language are acquired and integrated into a dynamic semantic system.
Universal semantics. Interest in the study of language meaning shifted from what keeps languages apart to what all languages are said to have in common. The idea that the meanings of words in different languages can be analyzed, at least partially, in terms of a given number of conceptual atoms identificable in the analysis of the vocabularies of all languages has become once again a very popular one with linguistics. As for the “universality” of grammar, it lies at the foundation of all linguistic work produced before the advent of structuralism. Linguistic and philosophical speculation ever since the 17th century has currently dealt with such problems. The current renewal of interest in language universals is due mainly to generative grammar which has always laid emphasis on those features which are shared by all languages alike.
The universalist point of view is based on the idea that language is basically an innate, or genetically inherited capability, which all human beings are “programmed” from birth to develop. This implies the adoption of the position that languages share the same basic conceptual framework. It can be argued that there is a universal set of semantic categories (i.e. categories concerned with time, place, causation, animacy, etc.) from which each language draws its own subset of categories, and it is only in the choice from this subset, and in the permitted combinations in which they are expressed, that languages differ.

The Child’s Acquisition of Conceptual Categories
How do we acquire conceptual categories in childhood? There are widely divergent points of view, extending from the empiricism of those who would argue that the cognitive system is learned entirely through experience from one’s environment (which includes cultural conditioning), and the extreme rationalism of those who would claim that the cognitive framework does not have to be learned, as it is part of an inherited mental apparatus specific to the human species. This polarity of views is obviously the universalist-relativist controversy in a slightly different guise.
Two prima facie arguments arising from modern linguistic research favour the universalist-rationalist point of view: as linguistics probes more deeply and precisely into the layers of linguistic structure, firstly it becomes more difficult to explain how a child learns so soon to manipulate the remarkable complexities of language, particularly on the semantic level, without having a “head-start” in the form of some fairly specific language-learning capacity; and secondly, it becomes easier to see how in a multi-layered analysis of language, widely different structures in phonology and syntax can be reconcided with identical, or al least similar, structures on the semantic level.
On the other hand, that at least part of concept learning runs according to empiricist thinking is clear from the way we observe young children to acquire the conceptual categories of their language by a procedure of trial-and-error. It has long been noted that learning a concept such as “cat” involves two complementary processes: (1) extension, i.e. extending the name one has learned to apply to same referents (cat1, cat2, cat3, etc.) to all objects sharing certain attributes of those referents (cat4, …. catn); and (2) differentiation, i.e. restricting the reference of a word to objects sharing certain characteristics, but not others (e.g. not applying the term cat to dogs, tigers, etc.). These two processes go hand in hand in the learning of category boundaries, but a child cannot learn both aspects simultaneously, so he tends either to overextend (e.g. “identifying “daddy” with all men) or to underextend (e.g. identifying “man” with all strange men wearing hats).

Creativity in Language
Discussion for and against semantic universals usually seems to assume that a language forms a static, closed conceptual system, and that once the fixed categories of the language have been acquired, our semantic equipment is complete. If this were true, it would cause us to take very seriously the sinister idea that our language is a mental strait -jacket, which determines our thought processes and our assumption about the universe.
But fortunately for the human race, language is only a mental straitjacket if we allow it to become one: the semantic system, like any other system relating to human society, is continually being extended and revised. In a language, new concepts are introduced in large numbers day by day and week by week, and in very little time, owing to modern mass communications, become familiar to many people. The technique by which the new concepts are introduced is lexical innovation, which may take the form of neologism and of transfer of meaning.
Language has within itself anti-creative pressures, and the function of the literary writer, in T. S. Eliot’s words, is to “purify the dialect of the tribe” – to restore the currency to its full value, and to resist the natural tendency to devaluation. Writers have always considered themselves the determined enemies of jargon and cliché.
Our linguistic competence (as Chomsky pointed out) is such that with a finite number of rules, we can generate and interpret an infinite number of sentences. Day by day we encounter and produce sentences we have never met in our whole life before. In its semantic aspect, this creativity of linguistic resource may be demonstrated by our ability to make up and make sense of configurations which have virtually a nil probability of occurring in day-to-day communication. But in performance, this creative or innovative power inherent in our language competence is eroded by our tendency to rely on well-worn paths through theoretically infinite array of possible English utterances. Thus not merely individual concepts, but configurations of concepts, become stereotyped; jargon invades syntax. The writer who resists this principle of least effort, by exploring new pathways and taking no meaning for granted, is in a real sense “creative”.
There is an important notion of linguistic creativity which applies pre-eminently to poetry: one which amounts to actually breaking through the conceptual bonds with which language constrains us. If one of the major roles of language is to reduce experience to order, to “prepackage” it for us, then the poet is the person who unties the string. It is in this context that the “irrational” or “counterlogical” character of poetry becomes explicable.
A very simple example of poetic irrationality in the Latin poet Catullus’ famous paradox Odi et amo: “I hate and I love”. The two-valued orientation of language makes us to see love and hate as mutually exclusive categories. But the poet, by presenting a seeming absurdity, shocks his reader into rearranging his categories; the stereotyped concept of love and hate as contrasting emotions is destroyed. A kind of conceptual fission and fusion takes place.
The quality just observed in poetic paradox is also present in metaphor – a more pervasive and important semantic feature of poetry. Again, the mechanism can be demonstrated by a very simple example. In an Anglo-Saxon poem, the expression mere-hengest (“sea-steed”) is used as a metaphor for “ship”. The connection between steed and ship lies in common connotations: both horses and ships convey men from one place to another; both are used (in the heroic context of the poem) for adventurous journeys and for warfare; both carry their riders with an up-and-down movement. By presenting the two concepts simultaneously, as superimposed images, the poet dissolves those linguistically crucial criteria which defines their separateness: the fact that a horse is animate whereas a ship is not; and the fact that a horse moves over land, whereas a ship moves over water. Metaphor is, actually, a conceptual reorganization. Through its power of realigning conceptual boundaries, metaphor can achieve a communicative effect which in a sense is “beyond language”. It has a liberating effect. As a chief instrument of the poet’s imagination, metaphor is the means by which he takes his revenge on language for the “stereotyped ideas” which have “prevailed over the truth”. (G. Leech 1990: 38). It is not surprising that children’s language produces many instances of semantic “mistakes” which strike the adult as poetic. G. Leech gave two of such instances: a child’s description of a viaduct as a window-bridge and of the moon as that shilling in the sky, both based, significantly, on visual analogy. The window-bridge example is very similar to the mere-hengest of the Anglo-Saxon poet: the openings in a viaduct, when seen side on, are indeed very close in appearance and construction to the window openings of a house. Using this generalizing ability, the child hits on physical appearance as a crucial criterion, at the expense of the criterion of function, which the language regards as more important. The difference between the two cases, of course, is that while the poet is familiar with the institutional categories and is aware of his departure from them, the child is not.

Conclusions
“Except for the immediate satisfaction of biological needs, man lives in a world not of things but of symbols” (General Systems Theory, p. 245). This statement by Ludwig von Bertalanffy is close enough to the truth to justify the concentration on the way language both determines and reflects our understanding of the world we live in.
Thinking of a language as providing its users with a system of conceptual categories, we may conclude:
1. That the concepts vary from language to language, and are sometimes arbitrary in the sense that they impose a structure which is not necessarily inherent in the data of experience.
2. That it is a matter for debate how for concepts vary from language to language, and how far it is possible to postulate semantic universals common to all human language.
3. That although the conceptual system of a language predisposes its users towards certain distinctions rather than others, the extent to which more is “enslaved” by his language in this respect is mitigated by various forces of creativity inherent in the system itself.

Bibliography:
1. Chiţoran, D. 1973. Elements of English Structural Semantics. Buc.: Ed. Didactică şi Pedagogică.
2. Leech, G. 1990. Semantics. The Study of Meaning. London: Penguin Books.
TOPICS FOR DISCUSSION
Comment on the two different conceptions in semantics relativism and universalism.

Chapter VI
SEMANTIC RELATIONS AND LEXICAL CATEGORIES

F. de Saussure directed the linguists’ attention to the necessity of studying the
multiple relationships among words in a systematic way. A particular lexeme may be simultaneously in a number of such relations, so the lexicon must be thought as a network rather than a listing of words. He suggested the existence of a network of associative fields, covering the entire vocabulary, and this structuring the huge mass of words. So an important organizational principle in the lexicon is the lexical field. This represents a group of words which belong to a particular activity or area of specialized knowledge, such as terms in cooking, sailing; the vocabulary of doctors, coal miners or mountain climbers. The effects are the use of different senses for a word and also the use of specialized terms. In fact, each word is a center of a ‘constellation’ or ‘series of constellations’, the point towards which other terms associated with it converge.
Saussure established four major types of associations among lexical items:
– etymological- based on resemblances in form and meaning;
– derivational- based on identity of affixes;
– semantic- based on meaning relations; formal- based on accidental form resemblances.
The types of associations listed above are illustrated by D. Chiţoran considering the example of the French word enseignement.
Enseignement

Enseigner apprentisage changement clement
Enseignons education armement justement

Progress in semantics was due mainly to lexicographic practice, which continued to bring together facts about meaning. Particular attention was paid to changes in the meaning of words. Traditional lexicology deals with types of lexical relations established considering distinctions similar to those belonging to Saussure’s conception:
– semantic ties – based on the signification of words; such ties result in synonymic and antonymic series of words;
– morpho-semantic ties obtaining among lexical items derived from a common basic element; they result in word families;
– syntagmatic ties obtaining among lexical items as they occur in actual utterance; syntagmatic ties may be divided into free – relations among sit and chair/ table/ down etc.- and stereotype – relations among lexical items part of set idioms and phrases, as a matter of fact, as mad as a hatter, day and night, etc.-;
– phonetic ties based on similarities of phonic substance; the first two examples represent minimal pairs, i. e. words which differ in just one phoneme, and the next two examples are words which present a common grammatical marker, in this case, that for the past participle
might – night flown- shown
town – down caught- taught.
These types of relations can be interpreted in terms of the distinction between expression- signifiant- and content- signifie-, as the interdependent planes of a linguistic sign. There are:
– formal or phonological relations established between the signifiants, i. e. the expression planes of linguistic signs; they account for homonymy;
– the relation of the type one signifiant- various signifies serves the designation of polysemy;
– the relation one signifie- various signifiants expresses synonymy;
– relations between various contents of linguistic signs.
E. Coşeriu pointed out that semantic relations should be signification relations, rather than relations between signs. Only in this way semantic structures can be distinguished from simple associative fields which are based on similarity relations between linguistic signs both on the expression and on the content level. The primary task of linguistics is to study the relational network encompassing the elements of language. The linguistic relational framework is structured along two axes, the syntagmatic and the paradigmatic axis. The linguistic elements situated on the syntagmatic axis find themselves in a both…and kind of relationship, i. e. they coexist within the same linguistic chain. They are in contrast position ( A. Martinet).The syntagmatic relations are directly observable in the spoken/ written chain. On the paradigmatic axis, the linguistic elements are mutually exclusive within one and the same linguistic sequence. They are in either…or relationship, in opposition (A. Martinet). These relations are not observable within a linguistic chain.
A. Paradigmatic Relations
1. The primary semantic relation on the paradigmatic axis is that of incompatibility, a relation which is characteristic of all lexical elements based on the substitution of items:
e. g. I had tea at breakfast.
I had coffee/cocoa/milk.
Part of the meaning of a term belonging to a lexical set is its compatibility with all the other members of the same lexical set in a given context. The wider concept of meaning incompatibility includes distinct types of oppositeness of meaning, each of them being designated by a separate term (J. Lyons).
a. Complementarity is a type of antonymic relation based on binary oppositions which do not allow for gradations between the extreme poles of a semantic axis; they are two- term sets of incompatible terms. Validity of one term implies denial of the other:
e. g. single – married
male – female
alive – dead.
b. Antonymy. The term is used to designate those meaning oppositions which admit certain gradations with regard to the meaning expressed:
e. g. young- old; young………childish/juvenil………….adolescent………….young………mature………middle……….. aged………..old…….ancient………
small- large; ….microscopic….tiny….little….small…..big/large…..spacious…..immense….
beautiful – ugly;
…..splendid…….wonderful….beautiful…..attractive…..handsome…..good-looking……. pretty…..nice….pleasant….acceptable……common…..ordinary…..plain…unattractive….ugly….horrible…awful….frightening….spooky….terrifying
clever – stupid;
interesting – boring;
fast – slow.
c. Reversibility refers to two terms which presuppose one another:
give- take; borrow- lend; buy- sell; husband- wife; offer- accept/refuse; employer-
employee. This type of binary opposition, a relation, involves a contrast of direction.
The relation can be realized by keeping the same lexical item and reversing the syntactic positions of the arguments:
e. g. John is the parent of James.
James is the parent of John.
or by keeping the syntactic positions of the arguments constant and changing the lexical form:
e. g. John is the parent of James.
John is the child of James.
Lexical pairs such as parent and child are called converses. Because of the alternative ways of expressing the same contrast, there arise cases of synonymy,
John is the parent of James = James is the child of John.
In case of ‘parenthood’ relation, the directional contrast is mutually exclusive, so there is an asymmetric relation.
Alf is parent of George. is incompatible with George is parent of Alf.
An example of symmetric relation is John is married to Susan. which entails Susan is married to John. In this case we talk about reciprocal relation.
d. Less common types of semantic opposition include hierarchic oppositions, which are multiple taxonomies, except that they include an element of ordering. Examples are sets of units of measurement- inch/ foot/ yard- , calendar units- month of the year- or the hierarchy of numbers which is an open- ended, that is it has no ‘highest’ term. The days of the week opposition is a cyclic type of hierarchy, because it has no first/ last member.
e. Last but not least, there is an interesting type of binary semantic contrast, called inverse opposition:
e. g. all – some willing- insist
possible – necessary still- already
allow – compel remain- become.
The main logical test for an inverse opposition is whether it obeys a special rule of synonymy which involves substituting one inverse term for another and changing the position of the a negative term in relation to the inverse term
e. g. Some countries have no coastline. = Not all countries have a coastline.
All of us are non- smokers. = Not any of us are smokers.
We were compelled to be non- smokers. = We were not allowed to be smokers.
It is possibly true that Jack is a hippy. = It is not necessarily true that Jack is a
hippy.
2. Another type of paradigmatic relation is synonymy. There are words which sound different, but have the same or nearly the same meaning. There is a tendency to limit synonymic status to those elements, which given the identity of their referential, can be used freely in a given context. There are no perfect synonyms, since no two elements can be used with the same statistic probability in absolutely all contexts in which any of them can appear. Synonymy is always related to context. Two lexical items are perfectly synonymous in a given context or in several contexts, but never in all contexts. The term used to describe this is relative synonymy. Context, that is the position on the syntagmatic axis, is essential for synonymy.
e. g. deep water *deep idea
profound idea *profound water
deep / profound sleep; deep / profound thought.
We can notice that the distinction concrete/ abstract is not relevant here, since words like idea and thought, both abstract, behave differently in relation to the pair of relative synonyms deep and profound. Talking about the terms used in describing synonymy, it is necessary at this point to present Lyons’ classification of synonyms into:
– absolute synonyms;
– partial synonyms;
– near synonyms.
Absolute synonyms should be fully, totally and completely synonymous.
i. Synonyms are fully synonymous if, and only if, all their meanings are identical ;
ii. synonyms are totally synonyms if and only if they are synonymous in all contexts;
iii. synonyms are completely synonymous if and only if they are identical on all relevant dimensions of meaning.
Absolute synonyms should satisfy all the three criteria above, whereas partial synonyms should satisfy at least one criterion (Lyons, 1981: 50-51).
D. A. Cruse (1987: 292) comments on Lyons’ classification, arguing that identical and synonymous are to be understood as completely synonymous; secondly, near- synonyms ‘more or less similar, but not identical in meaning’ qualify as incomplete synonyms, and therefore as partial synonyms, so the distinction between the two classes is not so clear as Lyons claims. Referring to absolute synonyms in language, Cruse states that there is no real motivation for their existence, and if they do exist, in time one of them would become obsolete, or would develop a difference in semantic function. For example, sofa and settee are absolute synonyms, but at a certain point in time sofa had the feature /elegant/, which now seems to have disappeared from the conscience of the speakers who use the two terms in free variation. But according to Cruse, this state of affairs would not persist, since it is against the tendency towards economy manifest in any language.
Examples like sofa and couch refer to the same type of object, and share most of their semantic properties-/ piece of furniture/ / used for sitting/ /with arms/ / backed/ / upholstered/-, so they can be considered synonymous. There are words that are neither synonyms nor near synonyms, yet they have many semantic properties in common. For example, man and boy imply /+male/ /+human/ features, but boy includes the property /+youth/, so it differs in meaning from man. The question to be asked is how to determine all relevant dimensions of meaning in order to establish the type of synonymy we are dealing with. Cruse draws a distinction between subordinate semantic traits and capital traits. Subordinate traits are those which have a role within the meaning of a word analogous to that of a modifier in a syntactic construction (e. g. red in a red hat).For instance, /walk/ is the capital trait of stroll, /good looking/ of pretty and handsome. For nag , /worthless/ is a subordinate trait.
Sometimes words that are ordinarily opposites can mean the same thing in a certain context, a good scare = a bad scare. The apparent synonymy of two utterances that contain a pair of antonyms hides opposite or at least different connotations.
e.. g. How old are you? – neutral connotation; inquiry about someone’s age
How young are you? You shouldn’t smoke. -negative connotation; it’s obvious you are too young to do that;
I don’t know how big his house is. – neutral
I don’t know how small his house is. -negative connotation; I know that it is too small
Even when using synonyms this implies not only a high degree of semantic overlap, but also a low degree of implicit contrastiveness,
e. g. He was murdered, or rather/ more exactly, executed.
He was cashiered, that is to say, dismissed.- the synonym is used as an explanation for another word.
Synonymy depends largely on other factors such as:
– register used, wife [neutral], spouse [formal, legal term], old lady [highly informal];
– collocation, big trouble *large trouble;
– connotation, notorious [negative], famous [positive]; immature [negative], young [positive].
– dialectal variations, which may be geographical ,- lift (British English), elevator (American English)-, temporal,- wireless became radio, -, and last but not least, social – toilet replaced lavatory, settee became sofa-,though the last two subtypes of variations cannot be always separated; (Cruse, 1987: 282-283)
– morpho- syntactic behavior,
e. g. He began/ started his speech with a quotation.
Tom tried to start/ *begin his car.
At the beginning/ *start of the world…
All the examples above refer to lexical synonymy, but there are also grammatical synonyms, operating at the level of morphology, means of expressing futurity, possibility, etc.
e. g. He will go / is going / is to go tomorrow.
He can/ may visit us next week if the weather is fine.
3. Hyponymy. Another type of paradigmatic relation is hyponymy / inclusion. It implies as a rule multiple taxonomies, a series of hypo-ordinate / subordinate terms being included in the area of a hyper-ordinate/ super-ordinate term. This relationship exists between two meanings if one componential formula contains all the features present in the other formula. Woman contains the features /+human/, /+adult/, /-male/.In different contexts, the emphasis is on one of the features included in the meaning of woman:
e. g. Stop treating me like a child. I’m a woman [= grown- up]
She is a woman [= human being], not an object.
She is a woman [ = female] , so she wouldn’t know what a man feels like in such a situation.
One way to describe hyponymy is in terms of genus and differentia. We can discuss about meaning inclusion, that is all the features of adult are included in woman, and about reference inclusion, that is all the objects denoted by woman are included into the larger category denoted by adult.
Sometimes we can’t have a super-ordinate term expressed just by one word:
musical instrument

clarinet guitar piano trumpet violin drums
B. Syntagmatic Relations.
Relations of the type both…and… are fundamental in structuring our utterances. The connection between paradigmatic and syntagmatic relations appears obvious, since in choosing a certain term from a synonymic series, we must take into account selectional restrictions. A particular type of arbitrary co- occurrence restrictions are collocational restrictions:
e. g. Ann/ The cat/ The plant died.
Ann/ *The cat/ *The plant kicked the bucket.
Collocational restrictions vary in the degree to which they can be specified in terms of required semantic traits. When fully specifiable, they may be described as systematic collocational restrictions:
e. g. Pass away /animate/ and kick the bucket /human/
Grill /meat/ and toast /bread/
When there are exceptions to the general tendency in collocating, we may speak of semi- systematic collocational restrictions:
e. g. Customer /acquiry of something material in exchange for money/
Client /acquiry of a certain type of service/, but a client of a bank is called customer, too.
The collocational ranges of some lexical items can only be dscribed by listing permissible collocants. Such items will be described as having idiosyncratic collocational restrictions. (Cruse,1987: 281)
unblemished spotless flawless immaculate impeccable
performance – – + + +
argument – – + – ?
complexion ? ? + – –
behavior – – – – +
kitchen – + – + –

The table above represents Cruse’s own intuitions. No semantic motivation can be discerned for the collocational patterns. It is debatable whether idiosyncratic restrictions are a matter of semantics at all.

Bibliography:
1. Chiţoran, D.1973. Elements of English Structural Semantics, Bucureşti: E.D.P.
2. Cruse, D.1987. Lexical Semantics,Cambridge: CUP.
3. Leech, G.1990. Semantics, London: Penguin Books.
4. Lyons, J.1977. Semantics, Cambridge: CUP.

Questions and exercises
1. Discuss the types of opposition relation.
2. Illustrate how various linguistic and extra- linguistic factors influence synonymy.
3. Match the appropriate adjectives with the nouns to show how collocation works.
calculated retirement
deliberate risk
voluntary judgement
premeditated mistake
considered murder
express ignorance
wilful wish.
4. Context is essential in choosing from a pair of synonyms. Think of contexts in which the following pairs of words cannot be interchanged:
hurry/ hasten pavement/ sidewalk
consider/ regard exit/ way out
injure/ damage spud/ potato
confess/ admit.
5. Synonymy and antonymy are associated when arranging words expressing different degrees of the same quality/ concept. The result is a cline/ scale. Try to arrange the following words according to their intensity:
a. immense, big, enormous, large, gigantic, spacious, colossal, extensive;
b. little, tiny, microscopic, small, minute, infinitesimal, diminutive;
c. distinguished, famous, well- known, illustrious, renowned;
d. mansion, castle, cottage, hut, house, palace, cabin.
6. A word can have different opposites in different contexts; which are they in case of:
Light bag/ wind/ colors;
Rough sea/ texture/ area/ person/ calculation;
High marks/ opinion/ building/ price/ temperature/ wind;
Hard exam/ chair/ journey/ work/ person/ drugs.
7. Construct hyponymy trees for vehicle, tomato, bench. Then complete diagrams like the following:
vehicle/ feature powered carries people four- wheeled
bus + + +
car … … …

Chapter VII
SEMANTIC THEORY WITHIN THE FRAMEWORK OF GENERATIVE-TRANSFORMATIONAL GRAMMAR

The development of generative-transformational grammar beginning with the late fifties of the 20th century has brought about a strong revival of interest in semantics. Particular mention should be made of the distinction postulated by generative grammar between deep structure and surface structure which is in many ways responsible for the recent developments in the study of language meaning.
Generative-transformational grammar resumes many of the concerns of traditional semantics. Thus, according to the theory, semantics should include an analysis of the way in which words and sentences are related to objects and processes in reality reintroducing into the discussion the problems of reference, denotation etc. Its second concern should be an analysis of the manner in which words and sentences are related to one another. These include an account of synonymy, antonymy entailment, contradiction, paraphrase, implication, presupposition, etc.

1. Semantics in the Standard Generative Theory of Language

A grammar of language can be described as a system of rules that express the correspondence between sound and meaning in the respective language. Every speaker possesses a finite and relatively small set of simple rules, which enable the speaker to produce and the listener to understand an infinite number of sentences. The set of rules represents – in the Chomskyan terminology – the linguistic competence, while the utterances produced on their basis constitute the linguistic performance.
Generative grammars are thus, synthetic models, able to generate all well-formed sentences in a language. By “synthetic” it is meant that starting from a set of rules arranged in a formalized construction, synthetic models lead finally to a set of utterances. G.T. is first of all, a model of competence, being-conceived as a model of language acquisition.
The rules are mainly of two kinds: rewriting rules and transformational rules. These rules are applied to symbols which make up the vocabulary of grammar.
Semantics will be concentrate on lexical categories and formatives (corresponding to words or full-lexical meaning or content words).
The organization of a generative grammar. Generative transformational grammar is defined in terms of 3 components: syntactic, semantic and phonological.
In the standard theory, the syntactic component in the most important one. It generates both the deep structure – which is semantically interpreted by the semantic component – and the surface structure which is further related to the sound aspect of language by means of the phonological component. While the semantic and the phonological components are purely interpretative, the syntactic component is basic to grammar since it represents the generative source of the grammar.
The syntactic component consists of a base syntactic subcomponent and a set of transformations, i.e. it has two kinds of rules: writing rules or phrase-structure rules and transformational rules. The first specify the form of constituent structure trees, and the second convert one kind of tree-structure into another (e.g. an active structure into a passive one). Transformations are rules that act on the phrase markers generated by the base, mapping deep structures onto the surface structures of sentences.
In the earliest published version of transformational grammar – Chomsky’s Syntactic Structures (1957) – meaning was in effect ignored. It was assumed that syntactic rules operated in complete independence from meaning: their function was to “generate” or specify by rule the grammatical sentences of a language, and to assign to these sentences their correct structure. In fact, many of the transformational rules, such as that which converted an active sentence structure into a passive sentence structure, happened in general to preserve the meaning of sentences unaltered (and therefore to be rules of paraphrase), but this was considered an irrelevant side-effect of such rules. However, after a pioneering article on semantics by Katz and Fodor (“The Structure of a Semantic Theory”, 1963), transformational grammar went through a period of conceding to semantics a more and more important position in linguistic theory. (Leech 1990: 343)
Scope and object of a semantic theory in generative-transformational grammar. A semantic theory describes and explains the interpretative competence of the speaker. This ability implies that a speaker can interpret sentences in the sense that he can relate them appropriately to “states, processes and objects in the universe” (Bierwisch 1971: 167).
A speaker can understand an infinite number of sentences, some of which he has never heard before. This is because he knows a number of rules on whose basis he can generate an infinite number of sentences. The rules are said to project a finite set or rules on an infinite set of sentences (Katz and Fodor, 1966: 481). The problem of formulating such rules represent the projection problem.
This problem requires for its solution rules which project the infinite set of sentences in a way which mirrors the way speakers understand novel sentences. In encountering a novel sentence, the speaker is not encountering new elements but only a novel combination of familiar elements. Since the set of sentences is infinite and each sentence is a different concatenation of morphems, the fact that a speaker can understand any sentence must mean that the way he understands sentences he has never previously encountered is compositional, i.e. it is based on his knowledge of the grammatical properties and the meaning of the morphems of the language. The rules the speaker knows enable him to determine the meaning of a novel sentence, by following the manner in which the parts of the sentence are composed to form wholes. As any speaker is able to grasp the difference in meaning between any two syntactically similar strings, this ability falling under the scope of semantic theory, it follows that the projection problem is fully solved only in as much as the grammar is supplemented by a semantic theory.
The aims and objectives of a semantic theory as part of the transformational-generative theory of language are:
a) to establish the meaning and the degree of ambiguity of a sentence;
b) to detect semantic anomalies;
c) to state the paraphrase relation between sentences;
d) to state other relevant semantic properties of sentences.
These objectives are self-evident for the innovative character of this semantic theory as compared to more traditional ones. While semanticists in the past were mainly concerned with the analysis of meaning (usually of isolated elements), the change in the evolution of meaning etc., the interest is now switched to the analysis of the meaning of sentences, and of their semantic properties. (Chitoran, 1973: 172).
The semantic component of generative-transformational grammar. The semantic component of a linguistic description is a projective device consisting of:
1) a dictionary that provides a meaning for each of the lexical items of the language;
2) a finite set of projection rules which assign a semantic interpretation to which string of formatives (or string of words) generated by the syntactic component. To arrive at a semantic interpretation it is necessary for each lexical item in a string of formatives to be assigned a meaning on the basis of the semantic information provided by the dictionary.
The projection rules then combine these meanings in a manner dictated by the syntactic description of the string to arrive at a characterization of the meaning of the whole string and of each of its constituents. This process reconstructs the way in which a speaker is able to obtain the meaning of a sentence from the meaning of its lexical items and its syntactic structure.
The dictionary part of the semantic component offers information on a lexical entry which is analyzed at four distinct levels.
At the first level, each lexical entry is categorized grammatically by indicating its syntactic marker, i.e. the grammatical class to which it belongs (noun, adjective, transitive, etc.). The semantic information proper, i.e. the specification of the meaning or meanings of the respective item is given under the form of semantic markers (as semantic categories of the type:  Animate,  Human,  Male, etc., which indicate the semantic relations obtaining among various lexical units and appearing therefore in the description of many of them) and distinguishers, which reflect the idiosyncretic elements in the meaning of lexical items.
Semantic markers and distinguishers are the transformational analogues of semes in the structural semantics (the first are similar to classemes and the second to semantemes). The distinction between semantic markers and distinguishers consists in the fact that semantic markers are used in the semantic description of more formatives (words), while distinguishers occur only in the description of a certain formative, individualizing it. For example in the case of the formative mammal the semantic marker is (+Animate) and the distinguisher is [they feed the young with their own milk]. The first can appear in the description of many formatives: mammal, fish, bird and the second is applied only to mammal. (E. Ionescu 1992: 192).
The fourth type of information provided by the dictionary refers to the combinatorial abilities of lexical items in a given syntactic construction to render a definite meaning. These rules of the combination of items in order to render a given meaning take the form of selectional restrictions in the dictionary suggested by Katz and Fodor. Thus, handy means clever with the hands when said of persons, and easy to use, convenient to handle when used of things and places.
The syntactic marker of an item is indicated by the grammatical terms denoting it; semantic markers are enclosed between normal brackets (…), distinguishers are enclosed between square brackets […] and selectional restrictions are given between angles ….
The second constituent of the theory is represented by the projection rules (amalgamation), whose object is to account for the semantic relations among morphems and the interraction between meaning and syntactic structure. Projection rules are ultimately responsible for assigning a semantic interpretation to a sentence.
This they do in the first place by associating to the lexical items of a given sentence S, those readings which are compatible with their syntactic categorization as revealed by the phrase marker of the respective S (Katz and Postal 1964: 18). The next operation that projection rules perform is to combine the readings of inferior constituents into derived readings of successively higher constituents until the readings for the whole sentence are arrived at. The process by means of which composite readings are arrived at by combining readings from each of the sets of readings dominated by a given node in a phrase marker, is called amalgamation. There is an interplay of syntactic and semantic relations in regulating the pairing of readings, since one condition for two items to be joined in syntactic relation, is that all selectional restrictions of one be included in the semantic markers of the other.
A closer analysis of the dictionary component of Katz and Fodor semantic theory reveals many similarities with previous approaches to the science of meaning. In fact, what Katz and Fodor do in their dictionary component of the theory is to rediscover the Aristotlean reference to genres and species (semantic markers and distinguishers) (Mounin 1972: 168).
As Coşeriu indicated (1968) what Katz and Fodor essentially do, is to study meaning along the semasiological direction, that is starting from a given signifiant, proper signifiés are assigned to it in a given context, following certain (syntactic) operations. In its original form the theory does not account for such well established facts as the existence of primary meanings and secondary ones, and in particular, it does not account for transferred meanings, and, in general, for the widespread use of metaphor in language.
An obvious criticism that was raised against the theory regards, as in the case of componential analysis, the very hypothesis according to which linguistic signification and semantic structure in general can be reduced to a relatively small set of “atoms” of meaning, with no residue whatever because this hypothesis is far from having been accepted unanimously (Chitoran 1973: 177).
2. Generative Semantics Versus Interpretive Semantics
The generative-interpretative controversy raged in the early seventies, but had no conclusive outcome. After a while the partisans of each side moved on the other topics of interest.
The popular labels generative semantics and interpretive semantics refer not so much to ways of studying semantics per se, as to ways of relating semantics to syntax. Both developed out of the Standard Theory of 1965 (Aspects of the Theory of Syntax) in wich a sentence was seen as organized syntactically on two chief levels: that of deep structure and that of surface structure. The surface structure of a sentence was derived from the deep structure by means of transformational rules involving such operations as the delition of constituents, the movement of constituents from one part of a sentence to another, etc. The rules which specified the DS were phrase structure rules, which spelt out the basic constituency of sentences in terms of categories like Noun Phrases, Verbs, etc. As it was previously mentioned, these rules made up the base component of syntax, and had as their output (after the insertion of lexical items) deep structures and the transformational rules made up the transformational component of syntax, and had as their output surface structures. Apart from syntax, which was the central part of the total grammar, these were two interpretive components: the phonological and the semantic. The phonetic interpretation of a sentence was derived from its surface structure by means of phonological rules, while the semantic interpretation of a sentence was derived from the deep structure through the operation of the so-called projection rules of semantics. The whole theory, therefore, through the interaction of its various components, provided a matching of phonetic outputs with semantic outputs (G. Leech 1990: 344). So, the theory provides an account of the pairing of meanings with sounds which any complete linguistic theory must attempt. The syntactic component has special status, being the point from which the derivation of both sounds and meaning originates. Among the special claims of Standard Theory are (1) that syntactic surface structure is the only level of syntax relevant to the specification of phonetic interpretation; and (2) that syntactic deep structure is the only level of syntax relevant to semantic interpretation. This second point brings with it the important principle that transformational rules are meaning-preserving; that is, they do not in any way alter the meaning of the structures that they operate on. This means, in effect, that all sentences that have the same deep structures have the same meanings.
We can see, Standard Theory provides for an interpretative semantic component; that is the meaning of a sentence is specified by the application of semantic rules to a syntactic base. It may be diagrammed as follows:
Standard Theory
Transformational Grammar 1965
Semantic Interpretation
(Projection Rules)
(Transformational Rules)
(Phonological Rules)
Phonetic Interpretation

Later, an important modification to the interpretivist position was proposed. Chomsky (1970), Jackendoff (1972), and others didn’t claim any more that all sentences with the same deep structures have the same meaning. Within this revised theory, deep structure reverted to being a level to be justified very largely on syntactic grounds alone.
Generative semantics like interpretative semantic, arose out of Standard Theory, but it developed along a quite different path. Lakoff, McCawley, Ross, and others, “deepened” the deep structure so as to make it closer to a representation of a sentence’s meaning, and they also “lengthened” the transformational process of derivation from deep to surface structure. Leech considers that the logical terminus of this process was reached (Ross and Lakoff 1967 and McCawley 1968) when the deep structure of a sentence was declared to be so “deep” as to be identical with its semantic representation. This now meant that base component, in the sense of Chomsky (1965), was no longer syntactic, but semantic. And since the deep structure was the semantic interpretation, there was no longer any need for the projection rules to supply an interpretation of deep structure. Projection rules therefore disappeared, and the resulting diagram was:
Generative Semantics Position
(Transformational Rules)

(Phonological Rules)
Phonetic Interpretation
Since it eliminates the projectional rule component, the generativist model has the advantage of overall simplicity of design. But, the simplification is necessarily at the cost of expanding the transformational component, and making the chain of transformational derivation for each sentence considerably longer than was envisaged by Chomsky in 1965. (G. Leech 1990: 347).
The generativists, in the main, stayed commited to the view that transformational rules do not change meaning. This proved the most vulnerable principle in their model, and was subject to the severest criticisms from interpretivists.
Within the framework of generative-transformational grammar, a “battle” is being fought not only between two rival semantic theories – interpretive semantics and generative semantics – but also between two versions of grammar: one which is syntactically based (the “standard” theory as developed by Chomsky, Katz, Fodor, Postal, including interpretive semantics) and another one which is semantically based (generative semantics).
In the standard theory, syntax is independent; it is the generative source of the grammar, which provides a deep and a surface syntactic structure. The deep structure provides all necessary information to the semantic component whose task is to assign semantic interpretations (readings) to the deep structures generated by the syntactic component.
With the generative semantics models, the semantic component is the generative source of the grammar. The semantic representations which initiate the derivation of sentences are independently generated, and are then mapped onto surface (syntactic) structures by means of transformations. (Chitoran 1973: 181).
Thus there have been two ways heading to generative semantics:
1. the revision of the standard model particularly of the notions of deep structure, selectional restrictions, etc.
2. a reappraisal of the semantic component, more specifically of semantic representation.
Leech (1990) considers that a simple way of defining interpretive and generative semantics is to say that in the one case the semantic representation of a sentence is derived from a syntactic base, whereas in the other, the (surface) syntactic representation is derived from a semantic base.
The same author proposes a three-component model of language (semantics-syntax-phonology) in which expression rules would have the function of translating (or “recoding”) semantic representations as syntactic representations, or vice versa (no directional precedence was assumed). Thus we have two separate bases, with syntax and semantics both having independent well-formedness conditions. In fact, various phonologists (Sampson 1970) have also argued for a phonological base. Hence, Leech’s model differs from both the generative and interpretative models in containing more than one base component (Leech 1990: 349; 351).
Bibliography:
1. Chiţoran, D. 1973. Elements of English Structural Semantics, Bucureşti: E.D.P.
2. Leech, G. 1990. Semantics. The Study of Meaning. London: Penguin Books.

TOPICS FOR DISCUSSION
1. What is the difference between semantic markers and distinguishers? Give some examples.

Chapter VIII
NEW SEMANTIC THEORIES

1. Categorization
The process of categorization is essential because it represents “the main way we make sense of experience” (G. Lakoff 1987: XI). This mental operation, which consists in putting together different things, is present in all our activities: thinking, perception, speaking etc. Categorization and categories are fundamental for the organization of human experience. Without this capacity of surpassing individual entities in order to reach a conceptual structure, the environment would be chaotic and forever new. (E. Cauzinille-Marmèche, D. Dubois, J. Mathieu, 1988).
Most of the concepts or mental representations correspond to certain categories and not to individual entities. Therefore, it is fundamental to know the mechanisms of categorization, trying to give an answer to the question: What are the criteria which decide that an entity belongs to a category? The objectivist current gives a clear answer: categorization is made on the basis of common characteristics. The experiential realism imposes a different view, based on prototype theory. G. Lakoff considers that the theory of prototype changed our conception about categorization, reasoning and other human capacities (G. Lakoff 1987: 7).
Necessary and Sufficient Conditions Model. One traditional approach to describing concepts is to define them by using sets of necessary and sufficient conditions. This approach comes from thinking about concepts as follows. If we have a concept like WOMAN, it must contain the information necessary to decide when something in the world is a woman or not. How can this information be organized?
Perhaps as a set of categoristics or attributes, i.e.:
X is a woman if and only if L
where L is a list of attributes, like:
X is human;
X is “adult”;
X is female, etc.
One can see these attributes as conditions: if something must have them to be a woman, then they can be called necessary conditions. In addition, if we can find the right set, so that just that set is enough to define a woman, then they can be called sufficient conditions; that means we have identified the right amount of information for the concept.
This theory views concepts as lists of bits of knowledge: the necessary and sufficient conditions for something to be an example of that concept.
The Aristotleian model of necessary and sufficient conditions, very largely used in philosophy, anthropology, psychology and linguistics is based on the following thesis:
1. Concepts and categories are entities with very clear borderlines.
2. The model is based on truth and false system: It is a dog provided that it fits the criterial conditions of the category “dog”.
3. The members of the same category have an equal status since each member has the features required by the definition of the category. So, each member is a good as any other.
One major problem with this approach has been that it seems to assume that if speakers share the same concept they will agree on the necessary and sufficient conditions: if something has them, it is an x; if not, not. But it has proved difficult to set these up even for nouns which identify concrete and natural kinds like dog or cat. Saeed (1997: 36) takes as an example the noun zebra. We might agree on some attributes: is an animal, has four legs; is striped, is a herbivore. The problem we face, though is: which of these is necessary? The first obviously, but the rest are more problematic. If we find in a herd of zebra, one that is pure white or black, we might still want to call it a zebra. Or if by some birth defect, a three-legged zebra comes into the world, it would still be a zebra. Similarly, if a single zebra got bored with a grass diet and started to include a few insects, would it cease to be a zebra? Of course, these seem rather whimsical or strange questions, perhaps problems for philosophers rather than linguists, and indeed this zebra example is just a version of Saul Kripke’s example about tigers (Kripke 1980) or Putnam’s fantasy about cats (Putnam 1962). Questions such as these have important consequences for our ideas about concepts: if we cannot establish a mutual definition of a concept, how can we use its linguistic label?
Another argument against necessary and sufficient conditions as the basis for linguistic concepts is Putnam’s (1975) observations about ignorance. Speakers often use words to refer knowing very little, and sometimes nothing, about the identifying characteristics of the referent. Putnam’s examples include the tree names beech and elm: like Putnam, many English speakers cannot distinguish between these two trees yet use the words regularly. Such a speaker would presumably be understood, and be speaking truthfully, if he said:
In the 1970s Dutch elm disease killed a huge number of British elms.
Perhaps as Putnam suggests, we rely on a belief that somewhere there are experts who do have such knowledge and can tell the difference between different species of trees. In any case it seems, as with other natural kind terms like gold or platinum, we can use the words without knowing very much about the referent. It seems unlikely then that a word is referring to a concept composed of a set of necessary and sufficient conditions, or what amounts to the same thing, a definition. The idea is that natural kind terms, like names are originally fixed by contact with examples of the kind. Thereafter, speakers may receive or borrow the word, without being exposed to the real thing, or knowing very much about its characteristics. As we have seen, philosophers like to use examples of metals like gold or silver. Any inability to identify correctly or define the substance silver does not prevent one from using the word silver. We assume that someone once had the ability or need to recognize the individual metal and that somewhere there are experts who can identify it empirically. Putnam speaks about a “division of labour” in a speech community: between “expert” and “folk” uses of a term. Only the expert or scientific uses of a word would ever be rigorous enough to support necessary and sufficient conditions, but speakers happily go on using the word.
The Prototype Theory. Because of problems with necessary and sufficient conditions, or definitions, several more sophisticated theories of concepts have been proposed. One influential proposal is due to Eleanor Rosch and her co-workers (Rosch 1973, 1975, Rosch and Mevis 1975, Rosch et al. 1976) who have suggested the notion of prototypes. This is a model of concepts which views them as structural so that there are central or typical members of a category, such as BIRD or FURNITURE, but then a shading off into less typical or peripheral members. So chair is a more central member of the category FURNITURE than lamp, for example. Or sparrow a more typical member of the category BIRD than penguin or ostrich. This approach seems to have been supported by Rosch’s experimental evidence: speakers tend to agree more readily on typical members than on less typical members; they come to mind more quickly, etc. Another result of this approach and similar work (e.g. Labov 1973) is that the boundaries between concepts can seem to speakers uncertain, or “fuzzy”, rather than clearly defined.
G. Kleiber (1999) speaks about two sciences of prototype theory: the standard theory and the extended theory. The standard theory corresponds to the period when E. Rosch and her team publish their work. According to prototype theory, the category is structured on two dimensions: the horizontal dimension (the internal structure) and the vertical dimension (intercategorial relations).
The Horizontal Dimension. The prototype is the best exemplar, the central instance of a category. This new conception is based on the following principles (Kleiber 1997: 51).
1. The category has an internal prototypical structure.
2. The borderlines of the categories or concepts are not very clearly delimited, they are vague.
3. Not all the members of a category present common characteristics; they are grouped together on the basis of the family resemblance.
4. An entity is a member of a certain category if it presents similarities with the prototype.
So, this approach allows for borderline uncertainty: an item in the world might bear some resemblance to two different prototypes. Here we might give examples of speakers being able to use the word whale, yet being unsure about whether a whale is a mammal or a fish. In the prototype theory of concepts, this might be explained by the fact that whales are not typical of the category MAMMAL, being far from the central prototype. At the same time, whales resemble prototypical fish in some characteristic features: they live underwater in the oceans, have fins, etc.
There are a number of interpretations of these typicality effects in the psychology literature: some researchers for example have argued that the central prototype is an abstraction. This abstraction might be a set of characteristic features to which we compare real items. These characteristic features of BIRD might describe a kind of average bird, small, perhaps, with wings, feathers, the ability to fly, etc. but of no particular species. Other researchers have proposed that we organize our categories by exemplars, memories of actual typical birds, say sparrows, pigeons and hawks, and we compute the likelihood of something we meet being a bird on the basis of comparison with these memories of real birds.
There is another approach to typicality effects within linguistics, which is interesting because of the light it sheds on the relationship between linguistic knowledge and encyclopedic knowledge. Kleiber called this approach the extended version of the prototype theory. Charles Fillmore (1982) and G. Lakoff (1987) both make similar claims that speakers have folk theories about the world, based on their experience and rooted in their culture. These theories are called frames by Fillmore and idealized cognitive models (ICMD) by Lakoff. They are not scientific theories or logically consistent definitions, but collections of cultural views. Fillmore gives an example of how these folk theories might work by using the word bachelor. It is clear that that some bachelors are more prototypical than others, with the Pope, for example, being far from prototypical. Fillmore and Lakoff (1987) suggests that there is a division of our knowledge about the word bachelor: part is a dictionary-type definition (“an unmarried man”) and part is an encyclopaedia-type entry of cultural knowledge about bachelorhood and marriage – the frame or ICM. The first we can call linguistic or semantic knowledge and the second real world or general knowledge. Their point is we only apply the word bachelor within a typical marriage ICM: a monogamous union between eligible people, typically involving romantic love, etc. It is this idealized model, a form of general knowledge, which governs our use of the word bachelor and restrains us from applying it to celibate priests, or people living in isolation like Tarzan living among apes in the jungle. In this view, when using a word involves combining semantic knowledge and encyclopaedia knowledge, and this interaction may result in typicality effects.
G. Leech (1990) considers that one of the flaws of the prototype semantics is that it reduces the role of conceptual semantics, in explaining word meaning, to the minimum of matching a word to a category. But the nominal view appears to be too restricted, because it can only be easily applied to common nouns (rather than to adjectives, verbs, etc.).
In addition to the category – recognizing ability, human beings also have a different order of cognitive ability – something which is much more closely tied to language – which is the ability to recognize structural relations between categories. (G. Leech 1990: 85).
Although the prototype theory was considered a veritable revolution, it is not a miraculous solution for all semantic problems and it cannot surpass all the difficulties which remain unsolved in the classical model of necessary and sufficient conditions. But, the theory brings three new elements of a great importance for lexical semantics.
(i) This theory allows us to integrate in the meaning of a word, characteristics excluded by the classical model, being considered unnecessary, encyclopaedic features;
(ii) It proves the existence of an internal organization of the category.
(iii) It also explains the hierarchical conceptual structure and intercategorial relations.
We also have to take into account that this theory is a theory of categorization, first intended for psychological goals.
The Vertical Dimension. Relations between Concepts. The relational nature of conceptual knowledge is an important issue in semantics. Words are in a network of semantic links with other words and it is reasonable to assume that conceptual structures are similarly linked.
Models of conceptual hierarchies are fundamental in the cognitive psychology literature. A model based on defining attributes was proposed by Collins and Quillian (1969). In this model, concepts are represented by nodes in a network, to which attributes can be attached and between which there are links.
Proponents of prototype theory, (Rosch et al. 1976) have also investigated conceptual hierarchies and have proposed that such hierarchies contain three levels of generality: a superordinate level, a basic level, and a subordinate level. The idea is that the levels differ in their balance between informativeness and usefulness. If we take one of Rosch et al.’s (1976) examples, that of furniture, the superordinate level is FURNITURE, which has relatively few characteristic features; the basic level would include concepts like CHAIR, which has more features, and the subordinate level would include concepts like ARMCHAIR, DININGCHAIR, etc., which have still more features and are thus more specific again. The basic level is identified as cognitively important; it is the level that is most used in everyday life; it is acquired first by children; in experiments it is at which adults spontaneously name objects; such objects are recognized more quickly in tests, and so on.
This model has proved to be very robust in the psychological literature, though the simple picture we have presented here needs some modifications. It seems that the relationship between the classic level and the intermediate term might vary somewhat from domain to domain: man-made categories like FURNITURE differ somewhat from natural kind terms, and the relationship may vary depending on the person’s experience of the categories. So a person’s expert knowledge of a domain might influence the relationship between the basic and subordinate levels. Tancka and Taylor (1991) suggest that experts on dogs and birds might have a different, richer structure at subordinate levels for these categories from the average person.
2. Cognitive Semantics
Toward the end of the 20th century, there is both a dissatisfaction with existing formal semantic theories and a wish to preserve insights from other semantic traditions. Cognitive semantics, the latest of the major trends which have dominated the last decades, attempts to do this by focusing on meaning as a cognitive phenomenon.
As is often the case with labels for theories, the term cognitive semantics might be objected to as being rather uninformative: in this instance because in many semantic approaches it is assumed that language is a mental faculty and that linguistic abilities are supported by special forms of knowledge. Hence, for many linguists semantics is necessarily a part of the inquiry into cognition. However, writers in the general approach called cognitive linguistics, and other scholars who are broadly in sympathy with them, share a particular view of linguistic knowledge. This view is that there is no separation of linguistic knowledge from general thinking or cognition. Contrary to the influential views of the philosopher Jerry Fodor or of Noam Chomsky, these scholars see linguistic behaviour as another part of the general cognitive abilities which allow learning, reasoning, etc. So perhaps we can take the label cognitive linguistics as representing the slogan “linguistic knowledge is part of general cognition”. (Saeed 1997: 299).
Cognitive linguists often point to a division between formal and functional approaches to language. Formal approaches, such as generative grammar are often associated with a certain view of language and cognition: that knowledge of linguistic structures and rules forms an antonomous module (faculty), independent of other mental processes of attention, memory and reasoning. This external view of an independent linguistic module is often combined with a view of internal modularity: that different levels of linguistic analysis, such as phonology, syntax and semantics, form independent modules.
Functionalism, with which cognitive linguists identify themselves, implies a quite different view of language: that externally, principles of language use embody more general cognitive principles; and internally, that explanation must cross boundaries between levels and analysis. Thus, it makes sense to look for principles shared across a range of cognitive domains. Similarly, it is argued that no adequate account of grammatical rules is possible without taking the meaning of elements into account.
This general difference of approach underlies specific positions taken by cognitive linguists on a number of issues: in each case their approach seeks to break down the abstractions and specializations characteristic of formalism. Studies in cognitive semantics have tented to blur, if not ignore, the commonly made distinctions between linguistic knowledge and encyclopaedic, real world knowledge and between literal and figurative language. Cognitive linguists consider that syntax can never be antonomous from semantics or pragmatics. So, the explanation of grammmatical patterns cannot be given in terms of abstract syntactic principles but only in terms of the speaker’s intended meaning in particular contexts of language use.
A further distinction that is reassessed in this framework is the traditional structuralist division between, to use Ferdinand de Saussure’s terms, diachronic (or historical) linguistics and synchronic linguistics. In his foundational lectures, de Saussure, attempting to free linguistics from etymological explanation, proposed his famous abstraction: a synchronic linguistics, where considerations of historical change might be ignored, as if in describing a language we could factor out or “freeze” time. This consideration has been accepted in many linguistic theories, but is currently questioned in functional approaches. Linguistic structures, in a functionalist perspective, have envolved through long periods of use and the processes of change are evident in and relevant to an understanding of the current use of language.
If we turn to meaning, a defining characteristic of cognitive semantics is the rejection of what is termed objectivist semantics. G. Lakoff (1988) assigns to objectivism the basic metaphysical belief that categories exist in objective reality, together with their properties and relations, independently of consciousness. Associated with this in the view that the symbols of language are meaningful because they are associated with these objective categories. This gives rise to a particular approach to semantics, Objectivist Semantics, which Lakoff characterizes under three “doctrines” (adapted from Lakoff 1988: 125-6):
a. The doctrine of truth-conditional meaning: Meaning is based on reference and truth.
b. The “correspondence theory” of truth: Truth consists in the correspondence between symbols and states of affairs in the world.
c. The doctrine of objective reference: There is an “objectively correct” way to associate symbols with things in the world.
In rejecting these views, cognitive semanticists place themselves in opposition to the formal semantics approach. For these writers, linguistic truth and falsity must be relative to the way an observer construes a situation, based on his or her conceptual framework. The real focus of investigation should, in this view, be these conceptual frameworks and how language use reflects them.
In the cognitive semantics literature meaning is based on conventionalized conceptual structures. Thus semantic structure, along with other cognitive domains, reflects the mental categories which people have formed from their experience of growing up and acting in the world. A number of conceptual structures and processes are identified in this literature but special attention is often given to metaphor.
Cognitive linguists agree with the proposal by G. Lakoff and Johnson (1980) that metaphor is an essential element in our categorization of the world and our thinking processes. Metaphor is seen as related to other fundamental structures such as image schemas, which provide a kind of basic conceptual framework derived from perception and bodily experience, and Fauconnier’s notion of mental spaces, which are mental structures which speakers set up to manipulate reference to entities. Cognitive linguists also investigate the conceptual processes which reveal the importance of the speaker’s construal of a scene.
A consequence of this view of language is that the study of semantics and linguistics must be an interdisciplinary activity. One result is that scholars working within this and related frameworks tend to stray across intra- and inter-disciplinary boundaries more easily than most. The approach to metaphor has been applied not only to the study of grammar and semantics, but also to historical linguistics, categories of thought, poetic language, rhetoric and ethics amongst other areas.
Bibliography:
1. Kleiber, G. 1990. La sémantique du prototype. Paris. Presses Universitaires de France.
2. Leech, G. 1990. Semantics. The Study of Meaning. London. Penguin Books.
3. Saeed, J. 1997. Semantics. Blackwell Publishers.

TOPICS FOR DISCUSSION
1. What is a prototype? Give examples.
2. What are conceptual hierarchies? Provide examples.
3. State the main principles of Cognitive Semantics.
FINAL TESTS AND QUESTIONS

1. Define semantics and its object.
2. The relation between semantics and semiotics.
3. Physei – thesei controversy.
4. Comment on the drawbacks of referential theory of meaning.
5. Apply the description theory of naming to the following proper names- for each name find two different descriptive sentences Karl Marx, New York, Jane Austen.
6. Give examples of situations in which the causal theory of naming functions. Can the descriptive and the causal theories of naming be combined?
7. Absolute motivation.
8. Relative motivation.
9. Find the archilexeme and the archisememe for the next series of words:
wallet, bag, case, purse, suitcase, knapsack .
10. Point out the advantages and drawbacks of componential analysis.
11. Define the notion of semantic field and state the main elements of the semantic field theory.
12. Linguistic Relativism versus Semantic Universals.
13. Two or more words may be close in meaning and yet not collocate with the same items. Which is correct:
The baby began to cry/ started to cry as soon as they had left.
I couldn’t begin / start my car; the battery was flat.
Before the world started, only God existed.
14. Words have distinct syntactic behaviour. Analyse the differences:
The plane leaves/ departs from Gatwick, not Heathrow.
We left the house at 6.
We *departed the house at 6.
15. Mark the following words with positive[+], negative[-] or neutral [n] connotation. If possible, try to establish relations of synonymy or antonymy between pairs of them
frugal, famous, extravagant, boast, generous, miserly, notorious, careful, brag, resolute, strict, advertise, obstinate, severe, praise.
16. What is the criterion that differenciates the following words belonging to the two series:
a. partner, colleague, ally, accomplice, comrade;
b. pal, mate, associate, companion, buddy, friend.
17. Correct the sentences if necessary:
There was a high difference between the two teams.
I am doing this exam because I want to achieve a step in my career.
His books commanded criticism from many people.
He had been found guilty of some slight crimes.
She won many competitions, forming fame in the process.
I was very grateful, because he had rescued my life.
18. What are the opposites of single, white, light, heavy. Provide contexts.
19. Identify the different types of oppositions:
The more the haste, the less the speed.
Marry in haste and repent in leisure.
If you lie upon roses when young, you’ll lie upon thorns when old.
Better to give than to take.
Spare when you are young and spend when you are old.
Faults are thick when love is thin.
A saint abroad and a devil at home.
Pride goes before and shame follows after.
An idle youth, a needy age.
This world is a comedy to those that think and a tragedy to those that feel.
A good beginning makes a good ending.
Unselfish parents have selfish children.
Promise little but do much.
A pair of lovers are like sunset and sunrise; there are such things every day, but we very seldom see them.
20. What are the elements of the semantic component of Generative- Transformational Grammar?
21. The organization of a generative grammar.
22. Generative Semantics versus Interpretive Semantics.
23. Explain the importance of the process of categorization.
24. Discuss the main theses of the two opposite models: the model of necessary and sufficient conditions and the prototype theory.
25. The vertical dimension of categories. Give examples.
26. Discuss the main theses of Cognitive Semantics.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: