Dimensions of Applied Linguistics
LEXICOGRAPHY:Principles and Problems in
Lexicology and Lexicography as
Developed In the Soviet Union

Prev | Home | Next

The prodigious activity in the area of vocabulary study and a keen interest in evolving a theoretical framework for making different types of dictionaries in the Soviet Union - these both indicate that lexicology and lexicography have been recognized by the linguists there as distinct disciplines of language study. The full-length treatment of lexicology even in their various textbooks (Agayan, 1959; Bulakhovsky, 1953; Muchnik, 1961; Reformatsky, 1955) attests to the fact that in Soviet linguistics lexicology exists as an autonomous level of linguistic analysis at par with phonology, morphology, etc. By contrast, no corresponding discipline has been recognized by the textbook writers of the West, and "to an American observer", to quote Weinreich (1963 : 61), "the strangest thing about Soviet lexicology is that it exists".

Among the many talented Soviet linguists, as a theoretician in the field of lexicology and as a devoted "dictionary-creator" in the area of lexicography, the name of the late L. V. Shcherba commands a high respect in the Soviet Union. His views on the various problems of lexicology and his approach towards creating various types of dictionaries (lexicography) are. responsible for establishing lexicology and lexicography as distinct scientific disciplines and, furthermore, for shaping critically the new directions in these areas.

Section I of this paper will be mainly devoted to some of the recurrent problems of lexicology and the insights shown by Shcherba in resolving them. This account may also be viewed as 2 modest contribution towards constructing a theoretical framework for lexicology, the basic tenets of which, though implicit in the writings of Shcherba, have not been explicitly stated and formally defined. The paper, thus, may also be taken as a rejoinder to the statement made by Weinriech (1963 : 66): "In the considerable body of Soviet literature on lexicology, it is surprisingly difficult to find a comprehensive statement of the theoretical foundations and research goals of the linguistic disciplines". Shcherba's approach towards linguistic study is the outcome of a fruitful blending of theory and practice. This has already been shown elsewhere (Srivastava, 1969; 1968) in respect of phonology and language learning. The same is true for lexicology and lexicography : he enriched his theories from his experience while working on the Russian (monolingual') dictionary - Slovar russkogo jazyka Academii Nauk - and creating a bilingual Russian-French dictionary; at the same time, "all of his lexicographic work is impregnated with the ideas connected with lexicology" (Istrina, 1951 87).

Section 2 is taken up with the elaboration of the basic typology of dictionaries and the methodological decisions dependent on this typology.

I

The following are the basic tenets underlying Shcherba's theoretical framework and research goals in the area of lexicology (though each of these postulates demands fuller exposition, we will confine ourselves for the present only to a summary statement).

Language is a dialectal unity of thought process and symbolic action and thus the concept of sign finds its full relevance only as a "bridge" between psychology and linguistics. This means that, according to Shcherba, language should be regarded not merely as a means of communication, but as a symbolic activity of thought process (compare this concept with that of Saussure who holds the view that in "linguistics proper" sign should be treated "in itself or for itself").

In language, thought process and symbolic action shade into each other and the overlap represents the verbal symbol, the basic unit of which is word. By "unit" Shcherba means that smallest product of analysis which keeps all the properties of the level of description of which it is a unit (1962; 1910 : 233). However, it should be noted that Shcherba holds that "the psychological primitive is the sentence - it is not that the sentence is constituted of words, but a word is born from the sentence through an analytic process (and, therefore, is not a scientific construct) and you know, it is born and (hence) it exists". The relationship between the thought process and the symbolic action has been represented analogically by the arrangement of knots and threads in a net (See, Fig.1). Note that the word (a knot) exists simultaneously on two dimensions: as a functional entity or an element of symbolism on the horizontal axis, and as a thought element on the vertical axis.

Though in language symbolic action and thought process interpenetrate so much that it becomes impossible to isolate them, ontogenetically they have different roots (Vygotsky, 1962 : 44). The consequence of the deep interpenetration is that language begins to function as a filter between the objective world and the speaker. It is for this reason that Shcherba (1962) wrote that "in every language the world is represented differently. We confuse things (items) with words - things are understood in the form in which they are given to us in words, and the supreme act of cultural development rests in setting 2 free the concepts from the imprisonment of words" .

chart


Relationship between thought and process and symbolic action

The whole field of language is divided into two - lexicon and grammar: the former is "the designation of autonomous objects of thought" while the latter can be understood as "the expression of relationship between the objects" (Shcherba, 1945 : 16). The level of grammar has four subsystems - phonetics, word-formation, form-formation, and syntax.

The word is a unit of the sentence and the sentence has two distinct aspects - one related to lexicon and the other to grammar. Therefore, a word may fall into two categories - lexical (word) and grammatical (word). Thus, "all the auxiliary words like prepositions, conjunctions, copula, some of the pronouns, many prepositional and conjunctive expressions exist only as grammatical items and, hence, totally disappear from the dictionary" (Shcherba, 1945 : 16).

The same item may be lexical in one context and grammatical in another. For example,

He was in America. (lexical)
He was an American. (grammatical)

Similarly, the Hindi word ja(na) can have two functions - lexical (to go) and grammatical (passive marker); rah(na) can have three functions - grammatical progressive aspect marker), lexical ('to live'), and lexical compounding as in intensifier (continuation of a state) (cf. Beskrovny, 1960).

Grammar deals with the rule-formation activity and the words generated through such an activity will always be grammatical words. Each language has enumerable ways of generating 'new' words. Contrary to this, the lexicon is concerned only with already 'made' words (Shcherba, 1945 : 17).

It should be asserted that if lexicology is a distinct level, its basic unit should also be distinct from the units of other levels. Lexical meanings are often confused with morphemic meanings. But morpheme is a unit of symbolic action, i.e., grammar (strictly speaking, a sublevel of grammar, i.e., morphology). It is altogether a different situation when a morpheme may exist as (lexical) word, and then it may have a lexical meaning as well. As all the units of grammar derive their significance by rule-function, morphemic meaning should also be defined in terms of 'rule of word usage'. On the contrary, lexical meaning is defined by the paradigmatic and syntagmatic relations between the autonomous objects of thought. The two types of meaning have been shown in Fig.2.

chart


How a lexical unit derives its semantics (psychological content) out of paradigmatic and syntagmatic relations can be demonstrated by taking the autonomous objects of thought (related to the 'mode of movement') of which a native speaker of Hindi is in possession. Fig.3 defines the semantic range of three Hindi lexical items - paedal 'on foot', gari 'kinds of transport on land', and jahaj 'kinds of transport in air or on water'. The lexical item gari can be qualified by many other lexical items, but all, of them must have a mode of transportation [-person, +land], e.g., m°tar- 'motor', rel- 'rail', gh°ra 'horse', bael - 'bullock', etc. On the other hand, jahaj can have only those which are [-person, -land], e.g., havai - 'air-' or pani (-ka) 'water-'. It is worth noting that the qualifying lexical items in the collocations are mutually exclusive.

(-person) paedal (+person)
jahaj(-land) gar?i (+land)

Fig.3
Semantic range of Hindi words - paedal, jahaj and gar?i

The upshot of the above discussion is that a distinction must be made between meaning as a sign-function and meaning as a lexical-function: the latter concerns a unit which is psycholinguistic, while the former defines the semantic component of a sign (Leontev, 1965).

By the very definition of 'lexical word' (lexeme) polysemous words are different words rather than one lexical unit with different meanings (see the discussion on the Russian word igla in Shcherba, 1940 : 70-73).

Similarly, Shcherba (1945 19) considers that truba 'trumpet' is a different word from its derivative trubac 'trumpet-player' because the latter designates a person while the former item designates an inanimate object. But truba and the derivative word formed out of the diminutive suffix -ock-a i.e., trubocka are the same lexical items. To take an example from Hindi, the pairs given in set A will be two different lexical items (as they refer to two different autonomous objects of thought) while the pairs in set B will be variants of the same lexical item.

A 1. cam 'leather' +suffix for 1. Camar
2. l°ha 'iron' profession = 2. luhar
3. s°na 'gold' -ar 3. sunar

B 1. am 'mango' +diminutive 1. amiya
2. khat 'cot' suffix = 2. khatiya
3. pul 'bridge' -iya 3. puliva

Word-formation may be realized through morphological, phonological (morphemic alternation), or word-compounding processes. It is in the treatrtient of compound words that Shcherba's contribution is considered significant.

Compound-words, as generally formed by the samasa-process in Sanskrit and German, are considered by Shcherba as words only on account o their form. Shcherba considers them essentially as simple units of 'parole' which he calls 'syntagm'. Most of the compound words in these languages are formed out of the speech process ('parole') and do not enter into the repertoire of 'langue' (Shcherba, 1945 : 9).

As opposed to such syntagms, there are compound words which are compound only in the historical perspective or sign-delineation; otherwise, synchronically they exist only as a simplex. For example, paro-xod 'steam-boat' (i.e., steamer), paro-voz 'steam-engine' (i.e., engine) are according to Shcherba simplexes. A parallel example from Hindi will be jala-vayu 'water-air' (i.e., climate).

Word-compounding elements may actually perform a discriminatory function in designating and locating the autonomous object of thought. By form they may be compound words; but by function they are lexical items. Such words have been classed by Shcherba as 'potential words' (1945 : 18). Russian examples - zeleznaja doroga 'ferrous way' (i.e., railway), zubnaja pasta 'dental paste' (i.e., tooth paste). English examples are black-berry, black-board; Hindi examples are kama-deva 'name of the god of love in Hindu mythology' and sivaratri 'name for a particular religious day'.

Word, as a unit of language, can be defined only in relation to its form and content. Shcherba's four types of words (mono-, bi-, potential-, and syntagm-) are thus, based on the different combinations of features which are binary in function and bilateral in dimension (See, Fig.4).
Phonetic Word
+
mono-word potential word
syntagm bi-word

+
Content Word
_

Fig.4
Word-types

In summary, it can be said that Shcherba's model presupposes the following theoretical foundations for lexicology -

1. Lexicology is a distinct level of linguistics and is directly opposed to grammar. (Compare this with the stand taken by Trager (1949 : 5) - "lexicology often passes out of the field of linguistics into metalinguistics ..." Similar is the view of Bloomfield (1933 : 274) who thinks that "the lexicon is really an appendix of the grammar

2. The basic unit of lexicology is the lexical word (lexeme) which is a unit in its own right, distinct from all other linguistic units. A lexeme may be defined as "the designation of the autonomous unit of thought". (Compare, this with Whorf's definition (1956 : 160) which equates lexeme with the 'stem'. According to the descriptivists like Trager (1958 : 5) it is only a list of morphemes, and for Hockett (1958 : 171) "a lexeme is always a grammatical form, by definition".)


3. As language is a dialectical unity of thought process and symbolic action, the researcher is confronted with the interpenetrating factors operating between two. it is for this reason that the concept of word as a unit of sentence structure moves from the lexical to the grammatical field and vice-versa. [Compare this stand with that of Wells (1947) who considers that a word is a 'heterogeneous' term and hence, it should be Excluded from linguistic description].

A word may be lexical or grammatical.

A word may he definable by its phonetic form alone (then it is a syntagm), or by its content alone (then it is a potential word), or by different opposite combinations of form and content (then it can be either a simplex or a complex word).
4) Similarly 'meaning' (the most abused term of the time) is also not, homogeneous. The model presupposes that a language has no homogeneous semantic basis (Zvegnitsev, 1957 : 93). There are two types of meaning: the psychological component of the meaning related to the thought process, and the sign component of the mean related to the symbolic action. [Compare this treatment with that of Hjelmslev (1953) or Antal (1953) who consider that "meaning is not a p;sychic phenomenon but that it is objective and does not primarily exist in the minds of the members of the speech community…" (Antal, 1953)].

5) By definition, polysemous words are not one lexical item with different meanings, but rather different 'autonomous objects of thought' assigned to different lexical items. In terms of the present-day discussion, the Shcherba model favours the stand of Weinreich (1966) rather than that of Katz and Fodor (1963) in regard to the concept of lexical item.
6) The psychological element of semantic content namely, lexical meaning, is like a field - "obsceje" (the universal - which is realized only through the individual. (For the concept of 'universal' and 'individual', see, Srivastava, 1959). The range of the lexical meaning is defined by the paradigmatic and syntagmatic opposition in which a given lexical item stands against all other lexical items of a language.

II


The cardinal point concerning dictionaries is that they are 'creation' rather than 'compilation'. Furthermore, lying behind them are the purpose and goals which determine a great many methodological decisions for reorganizing the data into a properly presented lexicon. As the purpose and goals for making dictionaries are always set against the practical needs of the user, the orientation of dictionaries differs from type to type, i.e., the choice in various types of dictionaries is related to the value-judgements arising out of the governing purpose.

Thus, three basic factors come to the forefront in evaluating any particular type of dictionary: (1) the purpose, (2) the user, and (3) the methodology. It is in this context that Shcherba wrote (1940 : 54) "Of course, one of the basic problems of lexicography is the problem of different types of dictionaries. It bears a direct practical significance and somehow the problem has always been empirically decided". In all, Shcherba talks about nine different types of dictionaries: (1) Reference (Vade-mecum) - RD, (2) Academic - Ad, (3) Encyclopaedic - EnD, (4) Thesaurus - ThD, (5) Explanatory - ExD, (6) Bilingual - BD, (7) Ideological - ID, (8) Synchronic SD, and (9) Diachronic - DD.

When opposed to the type ID, ExD and BD are classed in one group and the common group is termed as Common Dictionary - CD. The two types CD and ID are commonly grouped as General Dictionary - GD as opposed to EnD. In his paper entitled "An attempt towards the general theory of lexicography", Shcherba (1940) tries to unsnarl the theoretical base and premises on which different types of dictionaries are created. This he does through a series of dichotomies, i.e., (1) AD vs. RD, (2) EnD vs. GD, (3) ThD vs. CD, (4) ID vs. CD, (5) ExD vs. BD, and (6) SD vs. DD.

In this paper we propose to limit our discussion to exploring the different parameters and registers which underlie the different types of dictionaries and which, in a way, influence a great many decisions in respect of the methodology of lexicography. A schematic representation of the typology of dictionaries based on different parameters and registers is proposed (See, Fig.5). It should be mentioned here that though our attempt has been to explicate the parameters and registers envisaged by Shcherba, for the sake of presentation and further discussion, the order of parameters that has been followed does not represent the one Shcherba proposes.

2.1. Based on the concept of overall pattern and sub-systems, dictionaries may be classified as RD or AD. The overall pattern includes every lexical item that is in the repertory of a given language, productively or receptively. Here, the lexicographer accepts the complexities and 'heterogeneity' of a language in all its dimensions: temporal and spatial;

chart

social and cultural. RD is representative of that dictionary which has been characterized by Shcherba as a type "behind which does not lie any unified language consciousness: the collected words may belong to heterogeneous speech-groups of different periods and which do not in the least form a system" (1940 : 55).

Contrary to this, AD is based on the concept of 'coexistent systems'. Here, a lexicographer, by accepting certain defined parameters, consciously selects his data so that the lexicon of a language, regardless of its heterogeneity, turns out to be a self -contained and homogeneous system. This type of dictionary is characterized by Shcherba as a type "at the root of which lies a unified (real) language consciousness of a defined speech-group of a specified time-span" (1940 : 55).

Based on this parameter, these two types of dictionaries, according to Shcherba, reflect the following differences:

RD AD
a) Speech-groups are different hence, lexicon is heterogeneous.a) Speech-groups are defined hence lexicon is homogeneous
b) The aim is to know the meaning of one or the other words.b) The aim is to verify whether or not in this or that situation, this or that (already) known word can be used.
c) Descriptive in nature.c) Prescriptive in order to establish stylistic norms.

It is obvious that dialect dictionaries in general fall within the RD-type, but if the region is defined and that lexicon is selected in order to give a detailed picture of the regional lexicon, the dictionary will be considered to be of the AD-type. All technical dictionaries made for the use of the concerned professionals are of the AD-type.

It is in fact due to the lack of knowledge of the aims and objectives and, consequently the methodological principles following therefrom in respect of these two types of dictionaries, that the publication of Webster's Third New International Dictionary evoked a bitter response. It was characterised as "a massive and shameless display of ignorance of simple linguistic facts and of long established lexicographic principles" (Allen, 1964 : 433). To think that the lexicographer as a linguist-reporter is superior in kind to a linguistlegislator (Mathews, 1955) is also a fallacious stand because lexicography is an applied science and, like other applied sciences, its methodology is conditioned by its aims and objectives.

The second parameter drawn for organizing the lexical material is based on the nature of the relationship between the signifier and the signified: the relationship may be fluid or rigid. Where the relationship is fluid we find lexeme in the form of a 'word', and where it is rigid we have a 'term'.

"A word", to quote Vygotsky (1962 : 146), "means both more or less than the same word in isolation: more, because it acquires new content; less, because its meaning is limited and narrowed by the context". It is due to this fluid relation that the word mean in English, depending upon different uses, conveys eight different senses 'insignificant', 'cruel', 'intend', 'result in', 'signifies', explanation', 'implies', (Alston, 1964 : 10).

On the other hand, a lexeme becomes a 'term' when it shows a rigid relationship; it stands for a fixed notion of a given theory. "The status of term in the vocabulary is considered as dual: on the one hand, systematicity is manifested at its utmost in a term; on the other hand, a term enters a system of notions of a given sphere of terminology, which does not necessarily coincide with the system of the vocabulary of a given language. Therefore, a term is a 'servant of two masters' and cannot be a perfect example of a word regarded as a member of a lexical system" (Reformatsky, 1967 : 125).

Proper nouns are not 'terms' because the signified object is not defined against any specific theory, but they also, like terms, have a rigid relationship, i.e., they refer to fixed objects of the objective world (in contrast to a word which refers to an autonomous object of thought).

Thus, the dictionary which selects lexicon based on a rigid relationship between the signifier and the signified is of the EnD-type, while that which is concerned with the lexicon having a fluid relationship is of the GD-type.

Here, we would like to discuss in brief the solution that Shcherba offers for two of the questions which a lexicographer has to face during his work on both EnD and GD: (1) What should be the underlying principle for including or rejecting items of the dictionary of the EnD-type in the GD-type dictionary? (2) What should be the manner of offering meaning for 'terms' if they are selected as items for GD?

Shcherba's answer for the first problem is in favour of including in GD all those lexical items which, due to usage, lose their specificity. This happens because proper nouns may, at times, be' related to common nouns in the same way as a definite common noun is related to indefinite ones. For example, see how the semantic component of a meaning (of the word 'philosopher') is actualized in the two sentences given below:
1. A philosopher/philosophers is/are in the habit of appreciating form. (Common noun in a strict sense).

2. A/The philosopher went to the interlocutor. (Common noun referring to a specific person).

Shcherba argues that though Khlestakov is a name (proper noun) of a character of Gogol's comedy Revizor, his very name projects the image of an impudent liar and a flap; the name has lost its reference to an individualized character, so much so, that in Russian we get a derivative word like xlestakovscina. Similar words from Hindi would be ram, ravan, vibhisan, etc.

As regards the second problem, Shcherba is of the opinion that if a term in EnD should derive its meaning by the place it occupies in the theory of which it is a term, its semantic component in GD must flow from the social situation in which it is used by a native speaker in the overall pattern (and not as in a restricted code). Thus, to take Shcherba's own example, the manner of referring to the meaning of a term like straight line will be different for EnD (for geometry) and GD; in EnD if it is somewhat like 'the shortest distance between two points', in GD it will be explained as 'a line which deviates neither to the right nor to the left (and moves neither up nor down)'. It should be noted that the same 'term' - say, a term like particle may mean differently in EnD and GD, and the difference can even be noted in two EnD's if they refer to two different theories:

particle in GD 1. 'a minute piece or amount'
2. 'A minor part of speech; as an article (the), conjunction, etc.
(Webster's Coillegiate dictionary).
particle 1. 'marker' (EnDl - linguistics, Bloomfield, 1938 : 199).
2. 'a thing which has no precise position or velocity' (EnD2 - physics, F;ank, 1958).

The parameter on which dictionaries are classified as ThD and CD concerns really the distinction between what Sapir calls language and language system. The latter is Saussure's liangage', while his concept of language material can be understood in the sense of the term 'text' as used by Gardiner (1951 : 329): "It is true that the term 'text' is, in ordinary parlance, confined to what is written or printed, but for the purposes of a linguistic theory the term can and should be extended to anything that has been spoken or reported as spoken on some particular occasion ..."

Long before the transormational grammarians, Shcherba, as a linguist, put forth his laudable goal of finding out the native speaker's intuition in language description. But at the same time he laid stress on "differentiating the living language process which goes on in the consciousness of a native speaker from that of artificial linguistic procedures which are merely a convenient tool for certain operations carried out by the linguists themselves" (Srivastava, 1969 : 51).

According to Shcherba, it is on the basis of Oparolel (recevaja deiatelnost) and language material (jazykovyjmaterial) that a speaker deductively evolves his linguistic consciousness. (Saussure's 'faculte du langage' corresponds to Shcherba's recevaja sposobnost (Shcherba, 1931).

The distinction between the two types of dictionaries - ThD and CD, lies in the fact that the former is directed towards building and evolving linguistic 'intuition' by putting language material (as exhaustive as possible) at the disposal of a user (and hence, it insists even on listening hapax legomena) while the latter attempts at explicating the 'intuition'.

The following may be considered as the main differences between these two types of dictionaries:

ThD CD
a. Evolves 'intuition'.a. Explicates 'intuition'.
b. Meaning, from the point of view of the user, is assumed not defined.b. Meaning is assumed as defined.
c. Data-oriented, process-based.c. Model-oriented, explication based.
d. Deductive process in focus.d. Vade-mecum in focus.
e. Stress on the specific meaning of 'ailo-variants' (through which a speaker is supposed to know the universally.e. Stress on the 'universal' meaning of a lexical item with reference to which lailo-variants' are illustrated.
f. Citation are given for all those usages where a lexical item is capable of projecting a different image.f. Citations to illustrate the meaning.

The parameter on which dictionaries are divided into CD and ID types is the same that divides linguistic units into expression and content. it should always be borne in mind that in reality all linguistic units are composite wholes, and any such bifurcation is to be Laken as artificial. It is only for practical purposes and use that a linguist lays stress on one aspect or the other.

The main difference between CD and ID is that the former organizes its material on lexical items, defined first as 'phonetic word', while the latter accepts them first as 'content word'. This leads to the different treatment of polysemous words as a dictionary entry; CD, like Katz and Fodor (1963), considers them as one item with different meanings, while ID favours the approach of Weinreich (1963).

As CD is based on the opposition of words mainly on the expression level, phonetic forms demand the organization of the dictionary on alphabetic lines. Contrary to this, as concept words, are items of thought, and thought has different dimensions of realization, lexical items in ID are grouped into 'families' where each one of them stands for one particular psychological dimension. In a qualified sense it can be said then that the concept-words are, in fact, those lexical items which, being manifested in phonetic complex system, and this system is expressed in ID in such a way that they fall under one lexical item, but, in fact, are actualized by different signs, letters, etc. ' (Shcherba, 1940 : 78). In this respect it should be noted that dictionaries of synonyms are, in a sense, a sub-type of ID.

The parameter on which ExD and BD are placed in binary opposition is mainly conditioned by the different motivations and objectives with which a user consults the dictionary. Users of ExD are native speakers, and the target set for creating ExD aims at native speakers with a view to explain one or the other lexical item which might be half-known or totally un know to them. Users of BD are not native speakers of the language and they approach the language with the primary aim of understanding texts foreign to them by finding equivalents in their own language.

Shcherba's advice to lexicographers working on BD's in the Soviet Union has been the main driving force for shaping the new critical directions in the creation of good dictionaries of the BD-type. A summary statement will, therefore, not be out of place here.

(a) Remember that there are no two languages which have identical lexical items and semantic systems, and hence, in each language the world is represented differently. We conceive things in the form in which they are given to us in words.

(b) The so-called equivalent words across languages are never equivalent in semantic range. For example, the Russian word stol is translated in French by an equivalent word table, but the French word stands for the meaning 'desk (for writing)' or 'table (of figures)" while the Russian word designates 'desk (for writing)', '(cooking) board', and 'department' (Shcherba, 1940 : 65).

Thus, it is incumbent upon the maker of a BD to give the equivalent words not in isolation, but in a system in which the word stands (at least the over-lapping area must be covered). Following this line, Soviet linguist-translators (Gak and Lvin, 1962 : 19) enter in the French-Russian dictionary relationships in 'the manner given below:

A. education = 1, 4, 1. vospitanie = A,B
B. formation = 1, 2 2. podgotovka = B, C
C. instruct ion = 2, 3, 4, 5 3. obrazovanie = C, D
D. enseignement = 3, 5, 6 4. prosvescenie = A, C
E. apprentissage = 5 5. obucenie = C, D, E

In Shcherba's view, on account of practical difficulties, BD cannot offer real meaning and the exact semantic range for the words of a second language. It can (and so it does) only conjecture about the possible meaning in a given context.

Under such conditions, Shcherba advises the users of BD to avoid assuming the load of conjecturing and shift to ExD. According to him BD (in the form it exists today), is useful only for (beginners). A new - type of BD is, thus, necessary wherein the effort should be directed towards overcoming the limitations of BD by palliative measures. The new type of BD should have the primary objective of letting the user move in the thought domain of the second language by releasing their concepts from the imprisonment of words with which they are habitually associated.

BD, according to Shcherba, is a type where the lexicon must be supplemented by grammatical and stylistic apparatus as an index for the usage of lexical categories.

The last parameter which distinguishes SD from DD is relatively straightforward in so far as the theory of linguistics is concerned, but it is the most confusing one with reference to lexicography. The difference between SD and DD is conditioned by the temporal factor.

Shcherba raises a basic question - what kind of dictionary can be labelled DD? Should one consider a dictionary to be of the DD-type because it is based on the history of language and provides enough facts about the word-etymology of that language? Furthermore, can we consider a dictionary to be of the DD-type if it claims to be 'A dictionary of old-language upto xth century'?

For example, the purpose, as stated in the preface to The Oxford English Dictionary, originally entitled A New English Dictionary on Historical Principles, is as follows:

"The aim of this dictionary is to present in alphabetical series the words which have formed the English vocabulary from the time of the earliest record down to the present day, with all the relevant facts concerning their form, sense-history, pronunciation, and etymology". Can this dictionary be classed within the DD-type?

Shcherba's answer to this question is an emphatic no'. To cover every aspect of a language is the function of a RD. According to Shcherba DD is that type of dictionary which first of all, like AD, must define the temporal axis on which it wants to build up its lexicon, i.e., it must settle the starting point in the history of that language. Secondly it must register every lexical item which falls within that domain; thirdly, the listing should be exhaustive in the sense that it should include not only possible 'becoming' words and 'becoming' meanings but also the dead words- fourthly, it must reveal the changes in the form-content relationship existing within a word; and finally, it should reveal the movement of the lexical system itself, i.e., it -should reflect changes in the system as a whole.

NOTES

1. In the preface to his Russian-French dictionary (1939) Shcherba expresses his concern at the scornful attitude of Soviet linguists of the time towards lexicography.It is owing to this attitude and their deficient training in this field that, according to Shcherba, dictionary-making was till then confined only to the "compilation" (sostavlenie) activity. To him, dictionary making is a scientific activity and dictionaries are "creation" (sozdanie) rather than compilation. For creating an object, ... the nature and function of its constitutive elements must be worked out in detail, and consequently Shcherba (1940 : 72) was of the opinion that "any word which is even slightly complex in its nature must be a topic for a scientific monograph and, therefore, it is hard to expect to complete any good dictionary with speed".

2. Observe that this is not exactly equivalent to the Whorfian concept. Shcherba talks only about the word, conceding thereby, that at the sentence level of expression, any thought expressed in any language is translatable into any other.