An Introduction to Lexicography
Meaning and its Relationship to Form

Prev | Home | Next

Componential analysis

3.1Relationship of form and meaning:Meaning is the central and the most important concern of lexicography. The reader consults a dictionary primarily to know the meaning of a lexical unit. The entire work of a dictionary is oriented towards providing meanings of the lexical units in as clear and unambiguous a way as possible.

The word 'meaning' itself has different meanings. Ogden and Richards have enumerated 22 definitions of meaning in their work The Meaning of Meaning (186-187). Adam Schaff has collected various meanings of meaning for criticism (1962, 227). Meaning has been studied in many disciplines viz. philosophy, psychology, logic and linguistics from different points of view but no simple solutions to the problem as to what is meaning have been found out.

The branch of linguistics which studies meaning is Semantics or Semasiology. The relation between form and meaning, and the place of meaning in linguistic structure have been variously discussed by different works on the subject. There is no scope for discussing in detail all the theoretical aspects of semantics in this work. Our concern would be mainly to examine the relation between the form and the meaning, the different components of meaning, their interrelationships, the relation of meaning between different lexical units in the general frame work of the lexical system of a language, all from the point of view of their treatment in a dictionary.

The problem of meaning has been approached in different ways in linguistics, some of which are briefly examined here. The three more commonly known approaches or theories treat meaning as a thing, as an idea and as a behaviour.1

In the first theory the meaning of a word is treated as the thing named by it. Thing, here, refers to concrete objects like chair, actions like read and kill, states like understand and know, abstract notions like honesty and courage and qualities like green or red.

This theory has a number of limitations. It does not cover function and relation words like English no, Hindi se, kaa etc. nor does it account for such objects as phoenix, unicorn etc. which have only fictitious and mystic existence. Two expressions naming the same thing do not mean the same. e.g. the morning star and the evening star both refer to the same object 'the planet Venus' but have different meanings. An object may be referred to differently by different phrases at different occasions. The same woman may be referred to as a mother, a daughter, a wife by different persons. An apple may be called an apple or a fruit at different occasions.

The second theory, called the ideational or the mentalistic theory, postulates that the meaning of an expression is the idea or thought associated with the expression in the mind of the speaker or hearer. In this theory the notion of ideas is conceived as mental pictures and images. This accounts for objects like unicorn, phoenix not covered by the earlier theory. But this theory is also not free form shortcomings. The mental pictures and images change from person to person and from occasion to occasion. The same object may object may evoke different ideas at different occasions. table may be associated with a piece of furniture for writing at onetime and a present being given at the time of some function. This theory covers fairly satisfactorily the physical and concrete objects. But what about words for abstract notions like honesty, virtue, and chasity? These words might be associated with different mental images in different persons. Grammatical words like no, how do not produce any mental picture in the speaker-hearer. This theory also fails to cover a large number of sound forms or expressions.

The third theory, the behaviourl or the causal theory, holds that the meaning is not associated with ideas or mental pictures but with the manner in which a hearer responds to a word. In other words, the meaning of an expression is the stimulus that produces the utterance and the response it elicits from the hearer. This again fails to account for all the expressions. Even a nonsense expression may evoke stimulus and may get responses in the same way as a meaningful one. A hearer may respond to a stimulus in more than one way and all of them may be quite different from the one actually intended by the speaker. This theory again is not adequate to account for a correct meaning of meaning.                                      


3.2Referential or denotative meaning:The most widely accepted theory of meaning, based on the first, is the theory of abstraction or referential theory, also called the theory of signification. It is based on the conceptual or cognitive aspect of meaning. This theory recognises the following three components of meaning:-

(1) the sound form or the linguistic sign,
(2) the concept underlying the form, and
(3) the actual referent, that part of the extra linguistic world which the linguistic sign refers to.

These components have been given different names by different writers:
concept        sound form   referent
reference    reference     denotatum
signifie     signifier/significant  signified
designatum sign     
artha            abda or pada    vastu

According to this theory the relation between the sound form and referent is not direct. It is imputed and is mediated by concept or designation. Ogden and Richards have given the following basic or semiotic triangle to depict the way the meaning is related to the form and thing/referent.

thought or reference or concept
sound form or symbol                referent or thing

They remark, "between a thought and a symbol a causal relation holds, between the thought and referent there is also a relation more or less direct, between the symbol and the referent there is no relevant relation other than the indirect one (imputed relation)". (Ogden and Richards 1952, 10-11)

How is this concept formed? A person observes different objects of the extralinguistic world and marks different occurrences of the referent. He compares the similar and contrasting features of this referent with those of the other referents. This enables him to make abstractions and form a general idea or concept of the thing in his mind.

When a person sees a cat at the first occasion he may not know what to call it. But when he sees it again and again, he finds that it has certain characteristics, it mews, it drinks milk, it has a long tail etc. These characteristics of the 'referent' make the speaker form a concept of its picture in his mind. He compares them with similar other 'things' with the same characteristics. He understands that all referents having these characteristics are animals. The association of the linguistic form cat with the thing of the above description makes him understand that this particular animal is a cat. In formation of this idea he first observes the general features and then specific features. The former help him in locating it in a class and the latter to particular thing/referent, He observes that the features associated with cat are different from those of a dog, the latter barks, it is bigger in size etc.

We may take another example. When one sees a chair repeatedly he discovers certain qualities or features associated with it. He finds that it has four legs, it is used for sitting, it has a backrest and it usually (not always) has hands. He associates these features with the linguistic form chair by hearing it and concludes that all things having these characteristics are chairs. These features are different from those of a stool, a table, or a cot, all pieces of furniture, sharing some features with chair and not sharing some others. On the basis of these features the person forms the concept of the referent in his mind. These features are, thus, crucial or criterial for the formation of the concept. Every word has a bundle of such features which in their totality are crucial for the determination of the meaning of a word. It is not only the presence of these features which is crucial for the determination of the meaning even their absence is equally crucial. e.g. man and child both share some common features like being human. So this is a positive or plus feature. But whereas man has the feature of adult child does not have it. In the latter case adult is a negative or minus feature (See 3-11).
It is clear from the foregoing discussions that meaning is an abstraction or generalization based on the semantic feature of the events or things. The notion of abstraction is not confined to the objects which have actual and concrete referents in the extra-linguistic world only but extends to cover such fictitious and imaginary referents as unicorn, fairy etc. it is also applicable to abstract notions like honesty, virtue and actions as beat, kill etc.

Some other classes of lexical units like attitudinal symptoms like hallo, Hindi are, Skt. Hale 'o', pragmatic operators like please, deictic markers H. yah, 'this vah 'that' and abstract relations or relation words like Hindi meN 'in' English from etc. are also covered by the notion of abstraction. These lexical units, although they do not have actual referents in the extra-linguistic world, are comparable to them in that they either carry the information comparable to that conveyed by the denotative words, or they stand for them as yah 'this' vah 'that or denote the relational functions which make the communication possible. All these classes of lexical units have been stabilized as part of the language and have their own communicative value. The lexicographer includes and describes them in the dictionary although the mode and type of their definition and description is different from those of the denotational words.2

As the concepts are abstractions and generalizations, they are basically identical for all people at a period of time. But they are identified and described differently in different languages so that the features determining the meaning become language specific. What is crucial in one language may not be the same in another. What is denoted by hand and arm in English has only one word in Rusian ruka. Bhojpuri gor? denotes both the objects denoted by English leg and foot. The hole of the needle is described differently by different languages. "We speak of the eye of a needle, but the Kekehi Indians of Guatemala call it the face of the needle, the Lahu of South East Asia and the Piros of Peru speak of the nostril of the needle, the Haka Chins of Burma call it the mouth of the needle, the Tiddims of Burma refer it as the ear of the needle, the Milta Zapotecs in Mexico say the face of the needle and the Amuzgsos likewise of Mexico talk about the hole of the needle. (Nida 1958, 285)

The most notable lexical items in which languages differ in the organisation of the concepts into linguistic forms are kinship terms, terms for colour and words denoting time. The English word brother-in-law has seven equivalent in Hindi, bahanoii. 'sister's husband's big brother', devar 'husband's small brother', nandoii 'husband's sister's husband' and saad?huu 'wife's sister's husband'. Hindi bhaiyaa or daadaa 'elder brother' has the following words in Bengali denoting the order of the brothers according to age:

bar?daa 'eldest brother', mejdaa 'elder brother' (younger to the eldest)' sejdaa 'elder brother (younger to mejdaa)' raangaa daa 'elder brother 'elder brother' (younger to sejdaa) and chodaa 'elder brother' (immediately elder to the speaker). Oriya has only niila to denote both blue and green. English has eleven colour terms as compared to two in Jale, three in Tiv, six in Tamil and twelve in Angami. Garo has for English go and come a single verb re. Re indicates movement on somebody's part, but it does not indicate anything about the direction of the movement. English has no equivalent word for this. Garo raa covers areas of both bring and take. raa can often be translated by the English word carry although it cannot be used for a broader range of behaviour than carry. (Burling 1970, 10-11).

The word carry in many language of India does not nave an exact corresponding word in English. There are more than one words referring to the different modes and objects of carrying.

Ao Naga  apung    carry (a baby in a customary manner)
    apu          carry ( on shoulder)
  penlok  carry (on head)
  am    carry (in hand)

The same is true of cut, shut, blow, break, etc.  Coming to the notion of time we may compare the following terms dividing the twenty four hours of a day in Russian and Hungarian.


morning  utro   5 a.m. to 11 a.m.
day          den       12 noon to 4 p.m.
evening          vecher     5 p.m. to 11 p.m.
night      noch   12 midnight to 4

morning     peggel    6 a.m. to 8 a.m.
forenoon deleo     8 a.m. to 12
afternoon    delutian    12 to 6 p.m.
evening  este   6 p.m. to 10 p.m.
night        ejszaka  0 p.m. to 4 a.m.
dawn        hajlnae   4 a.m. to 6 a.m.

This is quite comparable to the concept of time for twelve hours from midnight to midday denoted by the following solar divinities in Sanskrit.
1. Aśvinsafter midnight, two parts the part in the dark is atmosphere and  the part in the light is the sun,
2. UÀas    Ð√vas ‘to shine) early dawn when it begins to be visible,
3. Suuryaa     the wife of Suurya, gradually getting more visible and more reddish,
4. Vr,Àaakapaayii The wife of Vr,s,aakapi, time not specified,
5. Saran,yaa    So called as it advances towards the sun, comes to an end with the rise of the sun,
6. TvaÀ¶r,  Temporal position not defined,
7. Savitr, when the sky, with its darkness dispelled is overspread by the  rays of the sun,
8. Bhaga previous to sun rise, followed by Suurya,
9. Suurya  Ðsr,,/su,/svir  all verbs of motion,
10. PuuÀan  .  sun becomes strengthened with the rays,
11. ViÀn,u    . full of rays. (Nirukta 12-1-18)

The features as described earlier which are crucial for the determination of the meaning are called semantic features or semantic components. Some of these components are basic, simple and common and are understood by all speakers of the language. These components are called the denotative components and the meaning based on them is called the denotative or referential meaning. This meaning forms the basic core of the meaning and makes communication possible.


3.3.Connotative meaning: but the denotative meaning alone cannot be adequate to describe the total meaning of all the lexical units. Let us compare the following sets of words:

Eng. 1. kill, murder, assassinate,

2. eat, partake, peck, wolf, devour,

3. Hindi marnaa, svargavaasii honaa, eN bolnaa,

4. khaanaa, liilnaa, bhakosnaa,

5. Bengali svaamii, bhaataar.

kill            .             . ‘put to death’ ‘cause to die’
murde+r            .             . ‘kill unlawfully and on purpose’
assassinate .  ‘to kill (especially an important politician, ruler) violently and treacherously especially for political reasons’.
eat            .            .  ‘to take into mouth and swallow it’
partake.            .  ‘to share the food’
peck            .            .  ‘to eat without appetite’.
wolf            .            .  ‘to eat quickly and greedily’.
devour            .            .  ‘to eat hungrily and greedily’

Hindi khaanaa             .  ‘to eat’

svargavaasii honaa  ‘to go to heaven’ ‘to die’.
eN bolnaa            .  ‘to die’ (vulgar)

Hindi khaanaa             .  ‘to eat’

liilnaa            .            .  ‘to devour’

bhakosnaa      .              ‘to wolf’

Bengali svaamii            .  ‘husband’
bhaataar         .         ‘husband’ (vulgar)

In the above sets of the words, we find that all the words of each set have some common semantic features, which give their basic or denotative meaning. Some of these viz. kill, eat, marnaa, khaanaa and svamii have only the common features or denotative meanings. We also see that the denotative meaning is the simple and factual statement about the object. It is colourless and neutral. kill means simply causing death, whereas in murder there are some additional or supplementary features which determine its meaning. In the same way partake, peck, wolf, devour, svargavaasii honaa, t?eNbolnaa and bhaataar have some additional features. These additional features add some extra meaning to the basic denotative meanings. These features are called connotation and the meanings based on them the connotative meaning or connotational component of meaning.

We are connotations added to words?when a speaker feels that the denotational meaning would not be enough to produce some desired additional effects like irony, joke, humour, euphemism and respect etc. he adds connotations to the words. The connotational component of meaning related to the emotive or expressive notion of the speaker or to stylistic values.

The connotational meanings differ from age to age, society to society and individual to individual. The word purdah does not have the same connotation now in a north Indian city as it had some years back. Honesty and virtue have different connotations in different societies. The following concept of 'good' in Guaica may be interestingly compared with a similar concept in any society: 'Good includes desirable food, killing enemies, chewing dope in moderation, putting fire on one's wife to teach her to obey and stealing from any person not belonging to the same band'. (Nida 1964, 79)

The word hospital has different meanings to an architect who builds it, to a patient who is treated there and to the person living near to the hospital who speaks of it as an identification mark for getting the address or route to his house.

The connotative meanings are peripheral and unstable as compared to the denotational meaning which is central and stable. It may be felt as new today and die tomorrow.

The connotative meanings are introduced into the language by individuals. Gradually they become socialized and become the part of the language. Originating with figurative extensions they are gradually transferred and are stabilized in the language. In the beginning these meanings are used occasionally but in course of time they attain frequency and are gradually used commonly. We may take the Oriya word maharii used for female attendants in temples. Originally meaning 'the female member of Mahar Caste' and the 'female servant of the god' the word now means something like a 'keep' or 'prostitute'.

Stylistic differences also come under connotations. When we compare the word pitaa 'father' paapaa we find that the former is stylistically neutral, whereas the latter is colloquial. A similar distinction can be made between friend and chum. The connotations include the following variations of language.

Slangs, professionalisms, jargonisms, vulgarisms, dialectal words, neologisms, formal, peotic, baby language etc.

The connotative meaning is vary significant for a lexicographer. When he analyses the semantic structure of individual lexical units, especially the polysemous ones, he has to distinguish between the denotative and connotative meanings for fixing the arrangement of meanings.   


3.4Collocational meaning:But the denotative and connotative meanings do not cover the total semantic peculiarities of lexical units. Let us examine the following: -

handsome and beautiful have the same meaning 'giving pleasure or delight to the mind or senses'. But they differ in their lexical valency, or potentiality of being combined or collected with other words. We say beautiful chair not *handsome chair. The words haath and paan?i in Hindi mean 'hand' but whereas paan?i grahan?a and haathmilaanaa are acceptable collocations *haathgrahan?a and *paan?imilaana are not.

Sanskrit maa and na are negative particles, maa is used with imperative klaaibyam maa sam gama paartha 'O partha! do not be unmanly' na is used with optative and benedctive.

In Bhojpuri the words nokar, carwaah and banihaar have the same denotative meaning of 'servant' but nokar is a general word for servant, carwaah is one who looks after cattle and banihaar is a ploughman.

In Halbi goNdh-, tulaa- and kaa- mean 'to cut'. But goNdh- is used for cutting firewood, tulaa for cutting vegetables and kaa fish and for other objects.

In the following words from Angami we find the same thing. Although the meanings are the same, the words have different ranges of application.

meta            . . to wash (hands)
kenu             .    . to wash (mouth)
chu             .      . to wash (face)
menyi             .        . to wash (clothes)
khru            .       . to wash (leg, body, head, utensils)

Some words can be collocated with only one set of words and not with others e.g.

Hindi baasii 'stale' can be collocated with roii 'bread' sabjii 'vegetable' but not with gehuuN 'wheat'. raman?iik 'beautiful' can be collocated with sthaan 'place' and not with lar?kii 'girl' or strii 'woman'.

In the above cases we find that besides the denotative and connotative components the collocation and range of application also govern the meaning of words. Collocation is also called co-occurrence or selective restriction.


3.5Grammatical meaning:The types or components of meaning described above are based on the interrelation of linguistic forms and the referents.
In girl, girls, girl's or khanaa 'eat' khaayaa 'ate' khaayegaa 'will eat' we find the same semantic components in all forms of the words.There is a recurring meaning identical in all forms of the words.This meaning is called the lexical meaning

Words do not occur in isolation. They are arranged in some patterns. There is some interrelationship between the linguistic units in these patterns denoted by morphological and syntactical devices. The meaning denoted by these devices is the functional or grammatical meaning.

When we examine lar?ke 'boys' striyaaN 'women', ghor?e 'horses' we find that these words have some lexical meanings, which are quite different for each of them. But they have some feature which is common in all of them. This feature is the function of plurality and the meaning denoted by this function is called the grammatical meaning. Certain classes of words, mainly the major parts of speech have both the lexical and grammatical meanings. The minor parts of speech e.g. articles, prepositions etc. have only grammatical or structural meaning. the words with lexical meaning are the content words or the lexical items and those with only grammatical meaning are function words or grammatical items.

The lexical meaning operates only at the level of word. The grammatical meaning, on the contrary, operates at different levels e.g. the morpheme level, the parts of speech level or in terms of grammatical function etc.

The number of grammatical items in a language is limited. Nothing is practically added to this class of words. The number of prepositions remains almost constant in a language. As there is little added to them, the grammatical items form the closed set in a language. On the other hand, the lexical items are indefinitely large. There is a constant increase in their number to fulfill the communication need of the speech community. The lexical set because of its being open to any addition belongs to the open set.

The lexicographer has to keep in view not only all the above components of the lexical meaning but also the grammatical meaning while defining his words.

3.6 Polysemy:
When one opens a dictionary he finds that most of the words have more than one meanings given either with or without numbers. e.g.


(1) a part of human body

(2) any opening etc.

 (1) to break,
(2) to pluck etc.
    (1) woman
    (2) wife

(1) empty,          

 (2) vacant etc.,

a common featured noted in the meanings of these different words is this that they are related to each other. Words having related meanings are called polysemous words and the phenomenon is called polysemy. The larger part of the vocabulary of a language consists of polysemous words. Monosemantic words are comparatively very few in number. They are generally scientific and technical terms, e.g

English .
. Hydrogen
'gas without colour, taste or smell, that combines with oxygen to form water.
'smallest unit into which a substance could be divided without a change in its chemical nature'.
alpapraan,a  'non-aspirate'

The more commonly used words are more polysemous than the less frequently used ones. Random House Dictionary College edition 1969 gives 134 meanings of the word run. The word tor?naa 'to break' has 16 meanings in Manak Hindi Kosh. Agni has 19 main meanings in the Sanskrit Dictionary (Poona). We may examine the different meanings of head here.

head of a person   . ‘a part of the body’.
head of a bed     . ‘upper end’
head of a com . ‘that side of the coin on which the head of    a ruler appears’
head of a cane . .   ‘top’
head of an organisation . ‘chief’, ‘position of command’
head of the cattle . ‘unit of a flock or a herd,
head of a glass of beer .‘foam on liquid that has been poured out’.  

We may also examine the different meanings of aaN kh ‘eye’ in the following.

aadmii kii aaNkh ‘the eye of the man’
aaNkh bacaanaa ‘to avoid attention’
suii kii aaNkh‘the hole of the needle’
ganne kii aaNkh ‘the sprout at the joint of a sugarcane’

In the same way the words hand, eye, ear, nose, tongue, leg, take, break, drink, etc.,  have more than one meanings. Not only the content words, the functional words are equally polysemous e.g.

English in has different meanings in the following a statue in marble, the man in top hat, women in green, bound in leather, written in ink, a teacher in him, inferior in physique etc.,

Hindi kii ‘of’ has different meanings in the following:-
Premchnd kii kitaab     ‘the book (written) by Premchand’
Pustakaalaya ki kitaab ‘the book in the library’
Mohan kii kitaab        ‘the book belonging to Mohan’
Raaj-kamal kii kitaab  ‘the book published by Raj-Kamal’.

From historical point of view polysemy is the change in the general semantic structure of a word. A word might have retained its original meaning and added some more meanings to it. It may also happen that some of the meanings have been lost and new ones are added. So when a lexicographer deals with a polysemous word he studies and analyses the interrelation and interdependence of its different meanings in the semantic structure of a word.

Now let us examine the following words: H. ghar has the following meanings: -

1. the place surrounded by wall with roof where people live, living place.
2. birth place, motherland, native-place.
3. family,
4. office,
5. room, apartment,
6. compartment,
7. case,
8. a place for accommodating anything,
9. main reason,
10. hole.

When we compare all these meanings we find that (1) is direct or central in an tie meanings. The other meanings are derived form it and are indirect or secondary having some additional semantic features. (2) is a semantic extension as the home of birth of which one is native i.e. the permanent or ancestral home, (3) is based on the notion of the inhabitants of a home who are related (may or may not live together), (4) is the place that houses an office. Other meanings are also related to (1) in the same way (9) is a figurative one, meaning house or abode of some event which when extended denotes the main reason (10) is used in combination will bat?an 'button' and actually signifies the house of a button.

Bengali kaal

1. time,
2. occasion,
3. age (as in aayukaal 'time of age')
4. death,
5. yama 'the god of death'

The relation between (1), (2) and (3) is clear. (2) and (3) are developed from (1), (4), denotes the end of the time of life, and (5) is one who causes it. Here again, we find that (1) is the direct or central meaning and other meanings are derived from it. We may examine some further examples and find out the relation-ship of meanings.

Tamil        munci      reedy sugarcane,
girdle formed of reed worn by Brahmin celibate  students
    Malto      akhunak      powder,
Malayalam     unaruka   to awake,
to understand, to know,
to watch,
to be vigilant,
to rise.

            Sanskrit kathaa
(1)   a tale, story,
(2)   a fable, feigned story,
(3)   an account, elusion,
(4)   talk, conversation, speech,
(5)   a variety of prose composition.

 Telugu aaku

(1)   leaf of a plant,
(2)   edible greens,
(3)   betel leaf,
(4)   leaf that is used as a plate to take food,
(5)   a palm leaf,
(6)   a page of manuscript,
(7)   an ear ornament,
(8)   a playing card, and
(9)   the spoke of cart wheel.

The meanings are related to each other (Reddy 1966. 72).

In all these cases we find that one meaning is central or direct. This meaning is the denotative meaning and has no supplementary or additional semantic features. The other meanings are derived and have additional semantic features.


Many words are used to denote these terms:
direct or denominative   indirect
main or primary       subsidiary or secondary
main or primary             peripheral
general              particular

All these words are used synonymously and are based on the logical relationship of the multiple meanings of a word. Other terms viz. 'dominant' and 'basic' or 'original' for central meaning are closely related to those noted above. The former is based on frequency parameter and the latter on the etymological and historical. They are also generally used for the earlier terms.

The determination of the primary or central or dominant meaning is a basic problem faced by the lexicographer. He has to fix them, interrelate the different meanings, find out the different subsenses characteristic of each meaning and arrange them in his dictionary in some systematic way. Different criteria may be used to find out the basic meaning.

One commonly used and widely accepted as reliable criterion is that of etymology. The meaning which is etymological or is nearest to the etymological is considered the basic or the central meaning: e.g.,

(1) fire,
(2) burning sensation,
(3) the fire of love
(4) love, affection,
(5) jealousy, hatred.

aag  is derived form Skt agni ‘fire’ and that is the central, basic and direct meaning. 


male parent,


founder or first leader



title used in personification etc.,

in these (1) is the etymological and the central meaning.

anga / aNg

(1)   body of human being esp. trunk as opposed to head (with its submenaings – breast of a woman, uterogenital canal, selfhood), 

(2)   nature of a person, aptitude, talent, flair, 

(3)   part, portion, branch, a specific direction, aspect, 

(4)   such participation (and other collocational sub-meanings) (Kelker, 1969, 53-64) 

Of these meanings, (1) is the central.  All others are extended and transferred from (1) which is the etymological meanings.  All the meanings of Skt. kathaa are derived from the (1) which is derived from kath ‘to tell’ ‘relate’.  But this is not always the situation.  Sometimes the etymological sense is relegated to a secondary meaning and the derived meaning attains the status of the central meaning. 


(1) good, pure,

(2) sharp,

(3) made crisp of baking or roasting,

(4) hard etc.

The word is derived form Sanskrit khara ‘sharp’ meaning (2) here


(1) an eunuch (esp. one employed to be in attendance in a royal harem),

(2) servant,    

(3) a respected man, chief,

(4) A caste of the Gujarati Muslims.

The word has its origin from khwaja (Persian) meaning ‘chief’ (3).

Such a determination of primary and secondary senses is possible only when the etymological and historical senses can be determined with the help of comparative reconstruction and historical attestation of the words.  This is possible generally with written languages only.  What about the following?

angami    keya        1. shelter,
  2. shade.
Ao Naga         ayung   1. call (to get attention)
2. invite.
  akok   1. win (in a contest),
2. succeed (in an attempt).
Khasi       bam   1. eat,
2. corode,
3. efface.

In a language where the etymological and historical studies have not been done it is impossible to apply this criterion. Even in the languages with written literature, many a time the historical sense is lost and the speaker ceases to discriminate between the basic and non-basic senses. The word crane in the meaning of (1) bird and (2) a machine is an oft quoted example. Sometimes the relatedness of even these words is not perceived by the native speaker. This leads to the development of homonymy.

Besides the etymological criterion, the lexicographer may apply contextual and statistical criteria to help him determine the central meaning.

The statistical criterion to determine the central meaning of words pertains to the frequency of its occurrence in the language. A frequency count of meanings has shown that the individual meanings of a word differ in frequency. The more frequently occurring meaning is taken as the basic meaning. The frequency of the word table in its meaning 'a piece of furniture' is highest (52%) of all the uses of this word; the meaning 'an orderly arrangement of facts' (as in census tables) is 35% and all other meanings have only 13% of use. Another example from the same source would make this point clearer. (West. 1959).

Room (space) takes less room, not enough room to turn
round.  (in) make room for (figurative), room for improvement.             12%

came to my room, bed room, sitting room, drawing
room, bath room (plural-suite, lodgings).  My room in coll-                  83%
ege, to let rooms .         .           .                                                    2%

The basic meaning is the one which occurs to us when we hear or see the word. For example when we hear the word head in isolation without any context the meaning which occurs to us instantly is 'part of body_______'. When we hear the word aag the meaning 'fire' comes first. This meaning is thought of as the representative meaning of the word without context by the majority of speakers. The secondary meanings occur only when used in a context.

In the Hindi examples aag, the meaning 'fire' is generally thought of by the speaker. The other meanings like 'burning sensation' occur only when we hear them in contexts like tumhari baatne mere sariir meN aag lagaadii 'Your words have produced burning (sensation) in me'.

puujaa (1) worship, adoration, veneration,
(2) bribe,
(3) punishment

Here, the first meaning occurs whenever one utters the word or hears it, the other two meaning are found only when used in such contexts as : piih puujaa karnaa; ‘to give punishment’;
usane apnaa kaam banaane ke liye saahab ko kaaphii puujaa car,haaii. ‘He gave big amount of bribe to the officer to get his work done’.

The central meaning occurs in various and vastly different contexts especially in different types of free collocations.  The secondary meanings occur in some contexts only.  Whereas table as ‘a piece of furniture’ occurs in many contexts and combinations, in the meaning of ‘the persons seated at a table’ it occurs only in a few contexts like ‘keep the table amused’ and in the meaning of ‘food provided at’ also in limited contexts like ‘he keeps a good table’.  Both of these meanings are not very common.

aag in the sense of ‘fire’ occurs in more contexts and has many sub-meanings like ‘hunger’ etc., but in the sense of ‘lust of love’ it occurs in limited contexts only.

Many vocables develop special senses for some technical languages.  The central meaning is used for the general sense and the derived meaning is used for the technical sense and for specialized contexts.

Eng.   bed    piece of furniture,
bottom of sea, a river, lake etc., layer of rock stone etc. (geography),
garden plot, piece of ground (for flowers, vegetables etc.,) as in seed-bed (geography)
Hindi        svar         voice,
vowel (grammar)
    ras    juice
sentiment (aesthetics)
Bengali       natthi-nathi   paper tagged by a thread,
certified paper (legal).

Another peculiarity of the central meaning is its use in the general language.  Special meaning is used for restricted language viz. slang, jargon etc.,

Eng. ass  donkey
buttocks (slang)
Hindi guru    teacher,
cunning fellow (restricted)
  Bengali      maal    a thing,
girl (slang),
wine (slang)
Kannada  tiirtha  water,
wine (slang)
Telugu ammavaaru  village goddess,
small pox (taboo)

Treatment of polysemy in a dictionary: How does a lexicographer find multiple meanings of a lexical unit and make sense discrimination among the different senses?  He analyses all the contexts in which a lexical unit occurs and by comparing the different usages similar and dissimilar in different contexts he extracts the different meanings.  for this he has to collect a fairly large selection of contexts.  e.g.

Hindi   (1) ab kyaa samy hE ‘what is the time now?’

(2) uskaa samay aajkal kharaab cal rahha hE‘Now a days he is having bad time’,

(3) samay aane par sab hiik ho jaayegaa ‘Everything will be  alright when time comes’,
(4) samay milne par meraa kaam kar diijiyega  ‘please do my work when you get time’.  

When the lexicographer compares all these contexts he finds that the word samay has been used in the following senses here:

(1)   time,

(2)   condition,

(3)   occasion,

(4)   leisure. 

Similarly from the following contexts, the lexicographer extracts the different meanings of ‘make’ noted against each sentence:

(1)   cloth is made from cotton. ‘manufacture’

(2)   why don’t you make yourself useful ‘cause to be or become’

(3)   make a living from ones’ writings  ‘earn’

(4)   they made me repeat the story ‘compel’, ‘force’, ‘cause’ 

We may examine the following usages of the Telugu word manci from which its different meanings are extracted.  

mancivaad,u          ‘a good fellow’

manciniil(u             ‘drinking water’

mancipuulu           ‘fresh flowers’

mancimanasu        ‘a kind heart’

mancisaruku         ‘good quality goods’

manciveela            ‘opportune time

manciuune             ‘gingely oil’

mancidi                 ‘yes’                              (Reddy 1966, 84) 

The determination of different meanings involves the analysing of all the contexts and all the combinatorial possibilities of a lexical unit.  The sum total of the meanings and sub-meanings of a lexical unit forms the semantic range of a word.  From this total semantic range the lexicographer examines the word in a particular context and finds out that in such and such situation a word has such meaning.  In other words the context individualizes the relevant meaning form the total semantic range of a word and eliminates the possibilities of other meanings.  Thus in usne apnii patnii ko patra likhaa ‘he wrote a letter to his wife’, the meaning ‘letter’ of the word patra is brought forth and other meanings viz. ‘leaf’ etc., are limited. 

Similarly from the following,  

he has paid the electric bill’ the meaning of bill ‘an account of money owed’ is abstracted and the other meanings viz. ‘written on printed notice’, ‘poster’, ‘placard’ and ‘proposed law to be discussed by a parliament’ are eliminated.  

The multiplicity of meanings is a source of ambiguity in a language.  The lexicographer faces this problem whenever he has to deal with polysemous words.  So he has to define his lexical unit in such a way that there is no possibility of its being interpreted in one way or the other. 

The ambiguity of meaning is removed by context which, as stated earlier, determines the relevant meaning in that particular situation.
That the context removes ambiguity of meaning has been stressed by many l8inguists and grammarians.  Bhar,tr,hari enumerates 15 contextual factors which help in determining meaning.  (Raja 1969, 48 ff).  he says that the vākya ‘the sentence-context’ and the prakaran,a “context of situation” help in disambiguating the polysemy of a word (Raja 1969, 53 also Reddy 1966, 72).  The role of context in determining the meaning of a word has been one of the basic principles of semantics in modern linguistics3.

The ambiguity created by polysemy is more frequently seen in what we may call depleted or incomplete contexts.  Let us examine the following:

(1)   It is a difficult case

(2)   He has gone to bring paper

(3)   usko iikaa acchaa lagtaa hE ‘the mark or an ornament suits her’.  

In these sentences the words case, paper and t(iikaa are polysemous.

Case has the following meanings (besides others)

(1)   instance or example of the occurrence of something, actual state of affairs.

(2)   (med.) person suffering from a disease,

(3)   (law) question to be decided in a law court.

Paper has the following meanings (besides others)

(1)   substance manufactured from wood, fibre, rags etc., in form of sheets used for writing, printing….,

(2)  newspaper,

(3)    a set of printed examination questions on a given subject.

t)iikaa has the following meanings

(1)   a mark on the forehead,

(2)   an ornament worn by ladies on the forehead.

The ambiguity in such sentences can be removed by quoting fuller contexts as the following:

(1)   The doctor said it is a difficult case,

(2)   He has gone to bring today’s paper,

(3)   usko sone ka t)iikaa acchaa lagtaa hE.  ‘golden ornament on forehead suits her’.

As a matter of fact that the major and the most important task of the lexicographer is to find out the different meanings of a word and present them in a dictionary.  In order to present the meaning in an umambiguous way he gives illustrative examples.

Different meanings of a word could also be discriminated by the interpretation of these meanings in terms of their synonyms in a dictionary. 

Bengali khuun blood.    lohit,rakta
murder,  hatyaa,badh
Hindi   samay time    vakta
period or epoch yug, kaal
condition     sthiti, haalar
leisure    phursat.
English   meeting assembly    congregation,gathering, convention, conference
duel   collision
slight    small    little, trivial, triffling, paltry,               insignificant.
slim    slender 
weak     feeble, frail, delicate
Marathi     gosta  [account of an incident, either fictitious or  fiction like in interest whether told informally] story, tale, kathaa, kahaanii, hakiikat, kissa,etc..
(2) the incident, the action, the fact in question ­­___,  (from Kelkar 1969. 57)  thing, matter, bhaag, prasanga, vastu, ciz. 
Sanskrit    naga hill,        parvata, giri, meru
tree  vr,kÀa, padapa
sun ravi, maartan,d,a, etc.

Another way of sense discrimination of polysemous words is by applying antonyms to the different senses of a polysemous word. 

Hindi kar,aa  (1) hard (antonym naram, mulaayam)
(2) difficult (antonym aasaan)
(3) cruel (antonym saday )
deep(1) going a long way down form the top (opposite of shallow)
(2) (of colour) strong, dark (opposite of light)
(3) profound (opposite of superficial)

While giving meanings to the lexical units, the lexicographer uses all the criteria described in chapter 5 for the sense discrimination of a polysemous word. He may not use all of them at a time an overt manner, but he has to keep a clear picture of all these in his mind.

But sometimes more than one sense appears to be dominant or central. e.g.

Hindi patra(1) leaf,
(2) written letter or document, note,
(3) newspaper,
(4) page.

Here both the meanings (1) and (2) appear to be central. Which of the two should be given first in arrangement of meanings and how should the subsenses be grouped? For the arrangement of meanings the basic meaning which is either etymological or is determined by frequency etc. may be given first, and the related senses may be grouped together on the basis of their logical relationship.

Of the five meanings of Bengali kaal, given earlier, (1) and (4) may be treated by some as almost equally dominant. In that case the lexicographer has to relate the different senses to these two. (2) and (3) go with the meaning (1), and (5) with (4). They are to be arranged in that order.

The word kalam has the following main meanings:

(1) pen,
(2) a painter's brush,
(3) a school or style of painting,
(4) graft,
(5) cutting or chopping.

Of these (2) and (3) will be related with (1) and (5) with (4).

The arrangement of these meanings is governed by the type of the dictionary.                                                 

3.7. Homonymy: in polysemy the different meanings of a word are related and have developed from one source. The lexicographer is faced with quite a different type of problem when he comes across the following situation:

The Hindi word man has the following meanings:

(1) mind; (2) desire, wish; (3) maund. Here we find that the meanings (1) and (2) are related, (2) is developed from (1) but the (3) can not be related to any of the two.

The first two meanings are etymologically connected, while the third is not.

The English word seal has the following meanings:
(1) kind of fish eating sea animal….,
(2) to hunt a seal,
(3) a piece of wax, lead etc., stamped with a design…
(4) something used instead of a seal e.g. a paper disc, stuck to, or an impression stamped on a document,
(5) to put a seal.

In these (1) and (2) are related in the sense that (2) is derived form (1). In the same way (3), (4) and (5) are related. (4) and (5) are developments of the (3). But there is no such relationship between (1) and (2) on the one hand and (3), (4) and (5) on the other. None is a development form either.

So can a lexicographer call them the different meanings of the same word? they are not related. Rather they are the different meaning of different words. Instead of treating man as one word with 3 meanings the lexicographer treats it as two words one with one meaning and the other with two e.g.
man2  maund.

Also seal.

seal1  n.(1) ‘a kind of fish eating sea animal’
(2) ‘to hunt a fish’
seal2 n.    (1) ‘piece of wax, lead etc., stamped with design’.
(2) ‘something used instead of a seal…’
(3) ‘to put a seal…’

We may examine the following words and see the same point:

English bark n. 'the noise made by a dog'; bark n. 'the skin of a tree'; bark n. 'a sailing ship'
Hindibar?aa n. 'a pluse cake': bar?aa adj. 'big'
Tamil maalai n. 'garland' : maalai n. 'evening'
Manipuritu pron. 'he' : tu n. 'ditch'
Bengali kaanaa n. 'one eyed'; kaanaa n. 'bank, side'
Telugu naad?u n. 'day', 'time', naad?u v. 'to enter' ; naad?u n. a country.
Malayalamaakalpam 'till the end of the world': aakalpam 'ornamental decoration'
Kuvi kaadi n. 'load of a cart', kaadi v. 'to scorch'
Ao Naga aki pp 'by', 'with'; aki v. 'own', 'possess'
Malto anda p. 'then' ; anda n. 'hunch of a bull'

A common thing found in these pairs of words is that their spellings and pronunciation are identical and the meanings are not related. Such words which are identical in spelling and pronunciation and have different unrelated meanings are called homonyms and the phenomenon is called homonymy.

But when we examine the following words we find a different situation: Eng. lead 'guide' or 'take' and lead 'soft heavy feasily melted metal…'; tear 'to pull sharply apart' and tear 'drop of the salty water coming form the eye…'. Although identical in spelling they are pronounced differently (li:d) and (led), (tEa) and (tia).

Such words are called homographs. Closely related to them are the words of the following type:
Right, write, rite; bye, by; reign, rain

Although different in spelling they are pronounced identically (rait), (bai), (rein).

Such words are called homophones.

When we compare all these three types of words described above, we find that the homonyms are words identical in both spelling and pronunciation and have different unrelated meanings; the homographs are words identical in spelling but different in pronunciation, these again having different meanings and homophones are words different in spelling but identical in pronunciation also having different meanings.

Homographs and homophones are not of much importance to the lexicographer, because they are quite different lexical units and pose no problem of relatedness of meaning or their arrangement. The homographs because of the identity of their spelling are treated like homonyms in the dictionary.4

This problem is rare in many Indian languages because of the syllabic script used by them.

While treating homonymy the lexicographer has to keep in view the different types of homonyms. In some cases there may be total identity in all the forms of a paradigm and its colocational possibilities. In others it may be only partial.

Let us consider the following examples:

English   seal ‘animal’    seal2   ‘wax…
pl. seals   seals 
possessive     seal’s         seal’s     
Hindi     sur1  ‘god’     sur2    ‘tune’
obl. pl.   suroN‘of gods’ suroN   ‘of tunes’

We find here that all the forms in the paradigms of the words seal1 and seal2, sur1 and sur2 are identical and from their forms one does not get any indication of their belonging to one word or the other. Such homonyms are called full homonyms or word homonyms. They are generally found in words belonging to the same parts of speech.

Another type of homonyms a lexicographer has to mark is found in the folloing cases:

Hindi                senaa n. ‘army’ senaa v. to hatch an egg

dir. pl. senaaeN ‘armies’                      sene    

obl. pl. senaaoN of ‘armies’   setaa    different inflected forms of senaa

sonaa n. ‘gold’                       sonaa v. ‘to sleep’

                                              sotaa,       different inflected

                                              sone  etc.    forms of sonaa

par ‘but’                               par ‘wing’                                             paroN        ‘wings’

In these examples only a few word forms are identical. Such homonyms are called partial homonyms or homonyms of word forms. In partial homonymy generally the canonical forms and some forms are identical and some other forms are not identical.

The lexicographer may meet some other cases of the following type which appear as partial homonyms:

Hindi                jiitaa                ‘alive’ (present participle of jiinaa to ‘live’),

                        jiitaa                ‘conquered’ (past tense of jiitnaa ‘to win’)

                        sone                 oblique form of sonaa ‘gold’,

                        sone                 an inflected form of sonaa ‘to sleep’.

                        sotaa               ‘a stream’

                        sotaa               ‘sleeping’ an inflected form of sonaa ‘to sleep’

The first two pairs of words do not present any difficulty because they are only some inflected and regular and predictable forms of the words. Since the lexicographer enters only canonical forms these words are not considered for the dictionary entry. But the last example is a little different. In this sotaa 'a stream' is itself a canonical form and deserves an entry. The identity of spelling of this form with the second word may tempt a lexicographer to enter the second one also a head word5. But when the forms are minutely examined such difficulty does not arise.

While treating homonyms, the lexicographer has to see if the difference in the meaning of these words is only lexical or also grammatical. When we compare Eng seal n. 'animal' and seal n. 'wax' and Hindi baal n. 'boy' and baal n. 'hair' we find that both the words belong to the same parts of speech. The difference lies only in their lexical meaning. Such homonyms are lexical homonyms. A lexicographer may be tempted to give these under one entry6.

When we compare such pairs as Hindi senaa n 'army' and senaa v. 'to hatch eggs' Malto ula adv. 'in day time', Ao Naga aki pp. 'by' and aki v. 'possess', we find that the difference is not only confined to the lexical meaning but the grammatical categories are also different. Such homonyms are called lexico-grammatical homonyms.

Some dictionaries7 treat polysemous words occurring in more than one parts of speech as homonymous and give separate entries for them. Such cases are called partial homonyms or grammatical homonyms. In these cases the words have the same canonical form but different paradigms and structural patterns e.g., verbs occurring both as transitive and intransitive, lexical units occurring as noun, verb, adjective etc.,

Eng.                 seal n. ‘wax…’: seal v. ‘to put a seal’

                        Drink v.: drink n.

                        Cut v. cut n. cut adj.

Angami             kekhre vi ‘to be folded’; kekhre vt ‘to fold’

                        Kekhre n. ‘fold’.

Abuj Madia      id adj. ‘this’: id pron. ‘this’

Hindi                gaanaa v. ‘to sing’ gaanaa n. ‘song’

Although etymologically related, such words are given in the dictionaries separate entries form the practical point of denoting the grammatical features and distributional characteristics of these words. This would be more helpful for a common reader especially for learning the generating power of the language.

The lexicographer should look at homonyms from the point of view of their origin. How do homonyms appear in a language? This will help the lexicographer in establishing the relatedness of meanings and distinguishing homonyms from polysemous words. The homonyms have the following sources in a language:

(1) Homonyms come in a language because of the convergent phonetic development of two or more different lexical units. Most of the homonyms in a language have their origin in this way. Phonetic changes pertaining to the loss of affixes, simplification of conjunct consonants, loss of syllables etc. affecting either one of the words or both of them lead to creation of homonymy, e.g.

Hindi kaam 'work' < Skt karma : kaam 'desire'
Sonaa 'gold'< Skt svarn?a: sonaa 'sleep'<Skt sup.

Punjabi sat 'powder'<Skt sattva: sat 'seven' <Skt sapta.

English sound n. 'strait' OE. Sund. 'swimming'
Sound 'healthy' OE. Zesund 'healthy'.

Homonyms of this types are in plenty in Pali and Prakrits.
Homonyms appearing by this process are etymologically unrelated. But in these cases also, even after the phonetic convergence, the two words may be taken as polysemous e.g., English ear 'organ of hearing' and ear 'spike of corn' are etymologically unconnected, but the two words are treated as related by some persons because the spike of the corn is considered as the ear of the corn in the same way as eye of needle, mouth or river, foot of hill etc. (Ullmann 1957, 128; Zgusta 1971; 77).

(2) Divergent sense development of polysemous words:

In course of time the different related meanings of a word become so very different that the relatedness of meanings is no longer perceived and the words are treated as homonyms. But this is very fluid and uncertain field. Much depends on the speaker's judgement which is not infrequently subjective. e.g.
OE. cest chest 'large box'
Chest 'part of human body'

The homonyms also have their origin in the language by borrowing form different languages. These borrowings may also undergo phonetic changes. The homonyms of this type may be of different types:

(a) One word is borrowed from another language: e.g.

Hindi                aam n.  ‘mango’ <Skt. Aamra

                        Aam adj. ‘general’ <Pers. aam ‘general

                        piir  n. ‘pain’ <Skt piid,aa

                        piir n. ‘a saint’ Pers. piir.

Bengali             tiir n. ‘bank’ Skt.

tiir n. ‘arrow’ Persain

Tamil                maalai n. ‘evening’ Dravidian

                        Maalai n. ‘garland’ Indo-Aryan

English mean ‘average’ Latin medianus

                        mean  ‘think’ OE. marnan

Telugu raaju ‘king’ < Skt. raajaa

                        raaju               ‘to kindle fire’ Telugu

(b) Sometimes both the words of the pair of homonyms are borrowed:

Tamil                kavi n. ‘poet’ < Skt. kavi

                        kavi n. ‘monkey’ < Skt. kapi

Hindi                kamaan ‘bow’ Persian

                        kamaan ‘order’  Eng. command

                        gallaa ‘noise’ Arabic. Gul

                        gallaa ‘crowd’  Pers. galla.

Malayalam        vaatam ‘argument’ < Skt. vaada

                        Vaatam ‘breeze’ < Skt. vaara

Bengali             aamin ‘an employee’  < Arabic amin

                        Aamin ‘amen’ < English amen

Sometimes when the number of homonyms is larger more than one words are borrowed.  e.g.

Malayalam        ari ‘enemy’ <Skt. ari

                        ari ‘lion’ < Skt. hari

                        ari ‘rice’ Malayalam

A particular type of homonyms found in Sanskrit are those formed by sabhangaśleÀas.  These words are split differently for different meanings.  These words are etymologically different.  e.g.

a-japa ‘one who does no worship’ aja-pa ‘protector of goats’

sa-mudra ‘with seal’ : samu-dra ‘sea’

suta-pa ‘the drinker of soma’ : su-tapa ‘practicing great austerity’8

One of the most controversial points in semantics having a direct bearing on lexicography has been the differentiation between polysemy and homonymy. The general principle of differentiation has been the relatedness of the meanings. If the meanings are related it is a case of polysemy, if not, it is a case of homonymy. The question of relatedness of meanings can be viewed form two points of view: historical and synchronic. Historically related meanings are those which can be traced back to the same source that is the meanings are etymologically connected or one meaning can be derived form the other. If they are not etymologically connected they are not related and are homonyms.

Synchronic consideration has the intuition of the speakers of the language as its basis. If the speakers perceive relationship between the meanings of a word it can be said that the meanings are related and form the semantic structure of one and the same word. if the speaker feels the meanings are not related it is a case of homonymy. But the intuition is not always a reliable guide. It can be quite subjective. The speakers may establish relatedness of meaning where there is none etymologically. We have noted the example of ear earlier.

The speaker may find the meanings unrelated in spite of the fact that these are derived form the same source. English crane 'bird' 'crane' a 'machine' may be treated as unrelated although the latter is a figurative extension of the same. Zgusta (1971, 85) In such cases the etymology is not relevant. An example from Nida may make it clearer. The form stock may occur in three very different types of contexts, e.g. he has a lot of stock in the warehouse, he sells stocks and bonds, he feeds the stock on his farm well. Though historically these three sets of meanings are related, for many present day speakers of English there seem to be no meaningful connections. (Nida, 1975, 13)

Similarly, the Hindi speaker may not find the two meanings of artha 'meaning' and 'wealth' related, although they are etymologically related. The two meanings of Hindi kar 'hand' and 'tax' may again be treated as unrelated.

In some cases the figurative extensions of the meaning may not be recognisable by the speakers e.g. in he will foot the bill the meaning of foot may not be related to foot 'the part of the body'. But if the meaning of foot which occurs in the foot of the column is taken into consideration a relation may be established in the sense that a person pays the sum at the foot of the bill. (Nida, 1975, 128)

The two meanings of the Tamil word miin 'fish' and 'star' may be treated differently by different speakers. Some speakers may not find any relationship between these two.

We observe form the above discussion that distinction between polysemy and homonymy is very uncertain and as observed by Lyons (1968, 406) is 'in the last resort indeterminate and arbitrary'.

What can a lexicographer do then? If etymological evidence is available he can depend on it. But in such cases where the etymological sense is lost he has to depend on the interpersonal impressions of the native speakers.

The etymological relationship can be established generally in respect of languages where some studies have been done. But in case of languages where such studies are not available the lexicographer has to depend either on the intuition of the native speaker or on his own intuitions.


3.8. Synonymy: In polysemy one word has more than one meaning. As opposed to this, there is a situation in which more than one words have the same or nearly the same meaning. Such words are called synonyms and the phenomenon is called synonymy or paryaava. Here, at times, different words which could have otherwise meant something else are used to denote one thing e.g. amara 'not dying', nirjara 'not aging' and vivudha 'learned' are used as epithets/synonyms of deva 'god'.

Indian languages have a long tradition of lexicographi work on synonyms. From Nighant?u to the present day there is a long history of dictionaries of synonyms. The origin of dictionary making itself may be traced back to the study of the synonyms Amarakosa has not only inspired many lexicographical works of this type in India, it has been used for preparation of dictionaries of foreign languages also. Roget's Thesaurus refers to this work.

Are there real synonyms in a language? This question has been widely discussed in books of semantics. Our concern here will be to deal with synonyms form the point of their sense discrimination and representation in a dictionary.

We may examine the following sets of words:
Hindi     pati    ‘husband’
svaamii     ‘husband (in sense of the master of the house)
bhataar    ‘husband’ (in pejorative sense used in abuses)
khasam  ‘husband’ (used as bhataar)
praan,anaath  ‘husband’ (the most beloved, in dialogues in intimate situation)
jiivansaathii  ‘husband’(companion of life, used in phrases like jiivansaathii kii talaas)
English    Attempt    implies making an essentially single effort.
Try      stresses effort or experiment made in the hope of testing or providing something.
Endeavour      hightens the implication of exertion and difficulty.
Essay    implies difficult but also suggests tentative trying or experimenting.
Strive     implies great exertion against great difficulty and specially suggests persistent effort.
 Bengali  strii   ‘wife’ (as in aamaar strii ‘my wife’)
mahilaa    ‘lady’ usually with bhadra (bhadromahilaa ‘gentle lady’) but not *aamaarmahilaa 
Skt. pr,thvii      ‘earth (in the sense of broad)
vasudhaa     ‘earth’ (in the sense of the place of jewels)
bhuumi     ‘earth’ (in the sense of substance)
Khasi kita     used for things little further off
kitei      used for things at a higher place
kitie    used for things lower or below. 
Kannada jaga         ‘used for any place’
niveÀana       ‘a site for building a house’
Halbi   ran       ‘very thick forest specially used when someone goes for hunting’.
ban        ‘nearby forest’.


When we analyse the semantic features of words of each set we find that there is no indentity in all the semantic features of all the words.  All the words of each set have some common feature (or a denotational meaning).  One word in the set has only this feature while others have some additional features also.  If two words have identity in all their semantic features they are called absolute synonyms.  The occurrence of absolute synonyms not only depends on the lexical meaning but also on the distributional characteristics of the words.  How to find out the absolute synonyms? 

One common test applied to know whether two lexical items are absolute synonyms or not is the test of interchangeability and substitution.  Two lexical units can be absolute synonyms if and only if they are interchangeable for one another in all contexts without the slightest change in their meaning.  But this is rarely found in a language.  Words similar in meaning may be substituted in some contexts but not in all contexts e.g. English exceptional and abnormal in the sense of ‘unusual’ can be interchanged for one another in the following sentence without any change in their meaning.

 The rainfall in April was abnormal / exceptional.  But in the following sentence:

My son is exceptional the substitution of exceptional by abnormal gives just the opposite meaning. 

Hindi lie, vaaste, hetu ‘for’ may be interchangeable in the following sentence:

Usne yah kitaab apnii lar,kii ke lie/vaaste/hetu khariidii hE. ‘He has bought this book for his daughter’.

But in the following sentences they cannot be interchanged:    

Vah kal kalkattaa ke liee (*vaaste, *hetu) jel gayaa.  ‘he went to jail for four months’.

jagha: sthaan ‘place’ may be interchanged in many contexts but in the following sentence they cannot be interchanged.  UnhoNne makaan ke lie jagah khariid lii hE.  ‘He has purchased a site for a house’.

This takes us to the field of selective restrictions or collocational potentialities of words.  Some synonyms can be collocated with some words and not with others.  Eng. broad: wide, broadest sense of a word is the same as the widest sense but we cannot say *wide transcription in place of broad transcription *wide idea in place of broad idea.   Again we can say country wide network and not *country broad net work. 

Bengali: khaas: viśeÀa ‘special’ can be interchangeable in khaas khabaar and viśeÀa khabar but khaas cannot be susbstituted by viśeÀa in khaasmahaal.

Kannada: gr,ahand mane ‘house’.  We have gr,hapraveÀa not *manepraveÀa and mane kat(t(u not *gr,ahakat(t(u   

Hindi: jaasuus and guptacar ‘spy’ are interchangeable in jaasuus / guptacar kaa kaam.  But we have jaasuusii upanyaas ‘spy novel’ and not *guptacarii upanyaas.

śava, laaś, murdaa, mr,tak mean ‘corpse’ but they cannot be interchanged in the following constructions:

śava (*laaś, * murdaa, * mr,tak)-daah, murdaa (*śava, * laaś, * mr,tak) ghaa/saraay. 

As a matter of fact, non-interchangeability of their components is one of the characteristic features of the set collocations.

Another thing to be considered by a lexicographer is the distribution-patterns of synonyms.  Even if they are interchangeable in some contexts, there is a difference in their syntactical valencies.  (Apresjan 1973, 181)

e.g. to answer : to reply

The former does not take a preposition but the latter takes one in the following:

to answer a question : to reply to a question

Anxious, uneasy and concerned have some common meaning, but anxious and uneasy take about whereas concerned takes at, for, with also besides about.

Hindi : adhik: bahut, ‘much’ bahut takes se where as adhik does not take it.  Adhik log          aaye hEN but bahut se log aye hEN

‘many people have come’

dekhnaa: taaknaa ‘to see’

dekhnaa :  takes both kii or and ko postpositions but taaknaa takes only kii or

usko/uskii or dekho ‘look at him’ and uskii or taako ‘behold him’,

but not *usko taako.

Besides this some synonyms have different paradigms, e.g. saphar and yaatraa refer to traveling form one place to another.  saphar emphasizes the distance and yaatraa implies destination.  Only yaatraa can be pluralized. (Bahl, 1974, 20).

Ahsaan, upkaar, saluuk, hit, hitsaadhan, bhalaa, bhallaii all refer to something done which is directly or indirectly beneficial to come one.  ahsaan forms Action Noun Phrase with post positions ke saath, ke uupar and par; upkaar with ke saath, par, ke prati and kaa; saluuk with ke saath; hit, hitsaadhan and bhalaa with kaa; and bhalaaii with ke saath and kii (Bahl 1974, 47). 

While analysing the synonyms the lexicographer would see that the principle of synonymy cannot be applied to polysemantic words.  Two words may be synonymous in some of their meanings but all the meanings of two polysemnatic words cannot be synonymous e.g.


kar and haath ‘hand’   

kar has the following meanings

(1)   hand,

(2)   the trunk of an elephant,

(3)   the rays of the sun.

haath means

(1) hand,
(2) measurement,
(3)   a bet.

We find that these words are synonymous in only one of their meanings.

The lexicographer would also note that all the meanings of a polysemantic word have different synonyms.


(1)   to see

(2)   to read-baaNcnaa, adhyayan karnaa

(3)   to enquire – muaaymaa karnaa

(4)   to search- khojnaa, talaaś karnaa

(5)   to examine-aajmaanaa, parkhanaa

(6)   to understand- socnaa, samajhnaa

(7)   to feel – bhognaa

Here we find that there is no synonymy between the synonyms of the different meanings of the word dekhnaa.

Bhognaa, samajhnaa, khojnaa and baaNcnaa are not synonyms.

We may compare the following set of synonyms to the five meanings of the word.  fresh (Arnold, 1973, 180).

fresh (as in fresh metaphor) fresh, original, novel, striking;

fresh (as in to begin a fresh paragraph) fresh, another, different, new;

fresh (as in a fresh air) fresh, pure, invigorating;

fresh (as in a fresh man) fresh, inexperienced, green, raw;

fresh (as in to be fresh with sub. ) fresh, impertinent, rude.

We find that there is no synonymy between rude, green, new, striking, the different meanings of fresh. 

The occurrence of complete synonyms in a language can further be tested by assigning antonyms to them.  The synonymous words have different antonyms for their different meanings;

English: firm and hard have the same meaning. 

            firm : opposite of loose (as in firm or loose decision)

            hard: opposite of soft (as in hard or soft words)

            deep and profound also have the same meaning

            deep: opposite of shallow (as in deep or shallow water)

            profound : opposite of superfluous (as in deep or superfluous knowledge)

Hindi: kahin and kar,aa ‘difficult’

             kahin : opposite of aasaan ‘easy’ (as in kahin or aasaan kaam)

             kar,aa: opposite of mulaayam (as in kar,aa or mulaayaum haath)

From the above discussion the lexicographer can find out that it is impossible to get complete synonyms in a language.  To quote Ullmann “absolute synonymy is an extremely rare occurrence, a luxury that a language can ill afford”.  (From Lyons 1968, 437).  

However, absolute synonyms are found only in a few monosemantic words which have technical meanings.  But here again the terms are not used by the same writer or school.  They are used by different schools e.g.  

Spirant                        : fricative

Flexion                       : inflexion

śailiivijn0aana         : riitivijn0aana ‘stylistics’

svanima                 : dhvanigraama ‘phoneme’

There are some very rare cases where we may find cases of absolute synonyms.  e.g. Hydrogen : H2, Vice Chancellor : VC.

But such cases are not found in the general language. 
So when a lexicographer deals with synonyms in his dictionary he actually describes or treats the partial synonyms or homoionyms.  In synonymy the lexicographer has to observe the overlapping of meanings.  That is to say that some meanings are identical in some of their semantic features and can be substituted for each other in some contexts and not in all the contexts.  

The lexicographer uses the synonyms mostly for defining lexical units.  Synonyms or semantically similar words are used for equating lexical units in a monolingual dictionary by a lexicographer.

We can examine the following sets of synonyms:
khaanaa, jiimnaa, bhakosnaa, liilnaa, huurnaa, bhojan karnaa.

khaanaa ‘to eat’ (used as a general word)
            jiimnaa ‘to eat’ (formal)
           bhakosanaa ‘to devour’
           liilnaa ‘to wolf’
          huurnaa ‘to eat with force’
          bhojan karna ‘to eat’ (formal)
We observe that in these words khaanaa is the only word which has a simple denotative meaning.  All the other words have this meaning as also some other additional connotative meanings.

In this set of synonyms the word with the denotative meaning is neutral.  From the point of view of its use, it is used more frequently.  It can be called the dominant synonym in the group.  Let us examine some more examples.  English hope, expectation and anticipation.  All these words mean having something in mind which is likely to happen.  Hope is not only a belief but a desire that some event would happen.  Expectation may be either of good or of evil.  Anticipation is, as a rule, a pleasurable expectation of something good.  In this group hope is the neutral word.  The words expectation and anticipation are used by educated speakers whereas hope is a word of general use.  So in this group hope is neutral and dominant.  (Arnold, 1973, 179) 

Hindi samay, kaal, vela meaning ‘time’

Samay has the general meaning of time; kaal is the extent of time from one point to another, vela is a definite point of part of time as in saandhyavelaa ‘evening time’, milankii vela ‘the time of meeting’.

In this set samay is used more frequently and in different contexts and is the dominant and neutral synonym. 
svarn,a, hema, hiran,ya, haaaka, loham, kanaka, kaan0cana meaning ‘gold’    
            svarn,a:                        gold in general
            hema :                          it has attractiveness
           hiran,ya/ haaaka :       it has yellow stuff
          loham :                        it has redness in colour
          kanaka :                      it is shining
         kaan0cana :                   it is fresh in colour 

Of these words svarn,a is more frequently used and is dominant or neutral of all the synonyms.

The lexicographer uses the dominant or the neutral synonym form the set to define or equate the non-dominant or non-neutral synonyms in his dictionary.

The synonyms can be discriminated in a dictionary by analysing their different semantic features and usages.  The following possibilities of discrimination of the synonyms9 may be utilized by the lexicographe

(1) One word is general another is specific:

Eng. refuse: reject.

Hindi khaalii : śuunya. ‘empty’

(2) One word has more emotive value than the other:

Eng.  politician : statesman

reject: decline


Marnaa : dehaanta honaa ‘to die’

(3) One word has a more polite expression than the other. 

Hindi dekhnaa: darśsan karnaa ‘to see’

Telugu tind,i : bhojanam ‘food’

(4) One word is used by one social class:

Hindi    arthii   :           janaajaa          ‘dead body’

            vrat     :           rojaa               ‘fast’

            kaakii  :           caacii               ‘aunt’

(5) In diglottic situations one word is used for the higher variety, the other for the lower:
             Bengali  manjan           :           maaNjan         ‘tooth powder’

Tamil    ilai                  :           ile                    ‘leaf’

            kud,dam           :           kud,o                ‘a vessel’

            veekamaka      :           veekamaa        ‘fast’

(6) One word is borrowed the other is native:

Hindi    telephon         :           duurbhaaÀ

red,io                :           akaaśvaan,ii

Telugu  bhojanam        :           tind,I     ‘food’

            aaim               :           samay  ‘time’

(7) One word is more colloquial than the other:

Hindi    skuul    :           paat(haśaalaa  ‘school’

            muus    :           cuuhaa            ‘mouse’

            saaNp  :           sarpa               ‘snake’

            paanii  :           jal                    ‘water’

Eng.     refuse  :           turn down                   

(8) One word may be more professional than the other:

Eng.     ophthalmologist          :           eye doctor

            Paedriatician              :           child specialist

(9) One word is tatsama and the other is tadbhava.

Hindi    karn,a   : kaan              ‘ear’

            kuÀ¶ha : kor,ha             ‘leprosy’

Telugu  adbhutamu ‘great wonder’, abhuram ‘wonder’

(10) One word is more literary than the other:

Eng.     passing away   :           death

Hindi    padhaarnaa     :           aanaa              ‘to come’

            lahnaa             :           paanaa            ‘to get’

            yaaminii          :           raat                 ‘night’

            bhojan             :           khaanaa          ‘food’

            śuunya             :           khaalii             ‘empty’

(11) One word may belong to child talk:

Eng.     daddy   :      father

Hindi    susu    :   peśaab             ‘urine’

Telugu  baayii  :     paalu               ‘milk’

            jijji     :    nidraa              ‘sleep’

Malayalam maamunnu:           coorunnu         ‘to have rice’

            Uvvaavu         :           muruva            ‘wound’

(12) One word may be local or dialectal:

Hindi    bEgan   :    bhanaa           ‘brinjal’

            aaaa    :   pisaan              ‘flower’

Eng.     lift        :   elevator           (American)

            flat     :  apartment       (American)

(13) One word may be more euphemistic than the other:

Telugu uppu is a taboo in the night so Skt.  lavan,am is used by some classes, lutta ‘basket’ is used by others; cavi ‘taste’ is also used.

Many lexical items related to human anatomy and activities related to them are treated as taboo in many societies.  Therefore, some other words are used for them.

Hindi    hagnaa   :    ta¶¶i karnaa    to secrete’

Telugu  dod,d,ikelu   : venakkellu`     ‘to answer the call of nature’

In the same way words relating to some diseases and death etc. are also taboo.  Other words are used in their place.

Hindi    cecak :    śiitalaa            ‘small pox’

            hEjaa    :  mahaamaarii   ‘cholera’.

These specifications will help the lexicographer to define the lexical units in a more clear way.

3.9 Antonymy: Another type of meaning relationship is what is traditionally called antonymy.  The term broadly signifies the oppositeness of meaning.  But it is open to criticism and allows different interpretations.  What is the antonym of kindcruel and unkind both can be good candidates for becoming antonyms of kind but unkind does not mean cruel.  Nor does not beautiful mean uglyGirl or man can be equally suitable candidates for becoming antonyms of woman.  The relationship denoted by antonymy may be analysed in terms of oppositions and contrasts in the meaning of the following types:-
(1) Extreme opposites or reversives or contradictories or binary opposites: this is the relationship of either, or i.e. the positive and negative features of the meaning.  e.g.

English.      dead    :   alive

                 tie     :     untie

Hindi    marnaa ‘die’ ; jiinaa ‘live’ 

These words show binary taxonomy or ‘absoluteness of boundary’ (Leech 1974, 106) and do not allow anything between them.  In other words, they are not gradable.  Dead means not alive.  There cannot be degrees of deadness and aliveness.

(2) opposites or polar opposites of contraries: this contrast can be explained in terms of scales running between two poles or extremes e.g. rich and poor.  good and bad, much and little.  Hindi t(han,d,aa ‘cold’ and garam ‘hot’, uupar ‘up’ and niice ‘down’.  One can be neither rich nor poor but not neither dead nor alive. There is a possibility of variability in terms of degrees between the opposites.  Between cold and hot there are cool and warmGood and bad can be judged form the degrees of goodness and badness.  In other words, these opposites allow grading.  Rich and poor may be judged from the degree of richness or poverty. 

(3) Incompatibles: this is the relation of meaning exclusiveness and not of contradiction.  Two meanings are incompatible if one contains at least one feature contrasting with a feature in the other.  e.g. red, black, blue, yellow, day and night, morning and evening.  To say day is to say not morning, or noon etc.  man and woman are incompatible because man has +male feature and woman does not have it.  Other meanings incompatible with woman are boy, cow etc. because they do not share features +adult and +human respectively.

(4) Conversives: this is the relationship that exists between words like English buy and sell, ask and reply, Hindi puuchnaa ‘ask’ and uttar denaa ‘reply’.

Besides these there are other relationships of meaning like ‘relational’ e.g. father and son, ‘hierarchical’ e.g. inch, foot, yard.

As in the case of synonyms complete or absolute antonyms are also rare in a language.  A word may have more than one antonym each restricted to some context in a language
The word thin has to antonyms (1) thick as in thin slice and thick slice (2) fat as in thin man and fat man.
The word dry may have the following antonyms, wet, moist and damp. eg.

dry lip              :           moist lip
dry air             :           damp air
dry clothes       :           wet clothes

 So the criteria of interchangeability cannot be applied for the antonyms also.  Good has poor as its antonym in some contexts e.g. good result and poor result and bad in others e.g. good boy and bad boy.

In lexicography the antonyms are used for the purposes of definitions.  The meaning of polysemous words can be disambiguated by antonyms.  The antonyms are also useful for discriminating the senses of polysemous words and synonyms.  (3.6 and 3.8).  The lexicographer should examine the semantics features of the antonyms while giving them as a part of the definition.

3.10. Hyponymy: another approach to the study of meaning relationship is in terms of meaning inclusion or hyponymy.  It relates to the analysis of the hierarchical relationship between meanings.  Many lexical units in a language include the meaning of other units in them.  For example vehicle includes in it the meanings of bus, car and motorcycle, animal includes in it the meanings of cat, dog, elephant etc.  The lexical unit whose meaning is included forms the lower rung of the hierarchical structure of meanings and is called hyponym (Sanskrit upa+naama).  The lexical unit in which this meaning is included is the superordinate term or hyperonym (Sanskrit upari+naama), sometimes referred to as classifier also.  The hyponym, some additional meanings which distinguished it from the class.  For example dog contains besides the meaning of animal also an additional meaning which distinguishes it from other animals like cow, cat, horse etc.  This hierarchical structure can be shown by the following diagram:


Another way to describe this relationship is in terms of genus and differentia.  The hyperonym is the genus or the general term.  The hyponym is the specific term.

Many of the terms in a hierarchical structure occur at both the levels-the hyperonym and the hyponym.  English dog is a hyponym of animal but a hyperonym for dog, bitch and pup.  In other words, it occurs both for the class as well as the specific.  We can illustrate this by the following diagram:


Here the word kuttaa is both a superordinate as well as an included term.

The hierarchical structure of meaning is used in lexicography for the definition of words.  A hyponym is defined in terms of a hyperonym e.g.

Bengali jonaakii ‘glow worm’ diiptiyuk%ta pokaa viśes$a ‘a particular shining worm’.
           English tiger ‘large fierce animal of the cat family……’

In the second definition there are two hyperonyms animal and cat from two levels of the hierarchical structure of the meanings. 

Taxonomy: the hierarchical structure of meaning described above is often referred to as taxonomy.  Taxonomy has been considered as a good method of organizing the description of meaning of the lexical units of semantic domains dealing with the phenomenon of the visible world.  It can also be used for description of the manner in which the mind perceives the outward reality.  One of the best known analysis of sets of terms into taxonomy has been Charles Frake’s treatment of the terminology of various skin diseases known to the Subanum, a tribe living in Mindanao in the Southern Phillipines. 

Specially notable has been the application of the principle of taxonomic structure in ethnosemantic studies in what is called folk taxonomies.  Applied to the study of a particular sub class of lexical domains displaying certain formal related properties it has shown that taxonomic structure is probably a universal characteristics of languages.  Even the most primitive languages have elaborate taxonomies to classify and describe objects of their environment.  Although there is difference in details of categorizations among languages there is no language which has an inferior system of categorization to the other.

 The principle of folk taxonomy can be used in lexicography, especially for primitive and unwritten languages, for clearly defining lexical units belonging to such sub-sets as flora and fauna etc.

3.11 Componential Analysis: we have seen that the meaning of a lexical unit is not an unanalysable whole.  It could be decomposed into its minimal distinctive features or components which contrast with other components.  For example the senses of the words man, woman, boy and girl might be expressed in terms of the presence or absence of certain features characterizing them.  These features can be written in a form of formula as given below:

+human= human                       +male= male

-human= animal                        -male= female

 So the senses of the above words could be expressed in the following way:

  humanmale      adult
woman         -   
boy            -  
girl            -        -  

The method of analysis called componential analysis has been mainly used for the purpose of describing sets of semantically related words.  It was evolved by anthropological linguists for the study of the relations in the kinship terminology10.  The relations in kinship have been analysed in terms of various features, like the following which account for their distinction.

(1)   Sex of the relative –male or female e.g. father and mother 
(2)   Generation of the relative with Ego ( ) = Ego’s generation; +1 = Ego’s parents generation; -1 = Ego’s children’s generation.
(3)   Consanguineal or blood relation vs. affinal relation or the relation by marriage.
(4)   Lineal or non lineal in a consanguineal relationship (father and son are lineal, uncle and nephew are not)
(5)   Degree of collaterality among non lineal relations based on the number of generations separating kinsmen from their common ancestors. (Southworth and Daswani 1974, 203-204)

These features are language and culture specific and may be more or less in number in different languages depending on the nature of the organisation of the kinship relationship in that language.

The componential analysis has also been applied extensively for the study of colour terms and pronouns. 

In lexicography componential analysis could be used for discrimination of synonyms (explaining of the overlapping of meanings), distinguishing the different senses of a polysemous word (by examining the additional semantic features and their links in the extended meanings) and the study of other relationships of meaning like antonymy, hyponymy etc. The application of componential analysis has some advantages.  Some unexpected features or distinctions in meaning are often discovered in the process of a through application of such a system.  It reveals the functioning of the system in its simplest form.

But this approach has certain limitations also.  It is applicable to a restricted set of words which contain the similar and contrasting features.  Componential analysis cannot be applicable to the analysis of the connotative and emotive meanings of the lexical unit because they analyse only the minimal features of distinctiveness.  (Southworth and Daswani 1974, 205).  For example the word bhaaii is used in the sense of ‘match’ or ‘mate’ in the following vah kumbakarn,a kaa bhaaii hE.  ‘He is a match for Kumbhakarna’.  This meaning may perhaps not be explained in terms of componential analysis so the componential analysis explains only part of meanings.

In spite of the short-comings, the componential analysis can be helpful in framing definitions in the hierarchical order of the components.  

Here it is worth while to make a distinction between marked and unmarked words in the same lexical set.  One word might be marked for some feature while the other might not be marked.  For example bitch is +female for sex and dog is not.

Mention may also be made of the classification of semantic features into semantic markers and semantic distinguishers as used by Katz and Fodor (1963, 170-210) for the analysis of polysemous words.  The semantic markers are features present also in the lexical meanings of other words.  They put the lexical unit in a class.  The distinguishers, on the other hand, are individual features not occurring in the lexical meanings of other words.  The distinguisher particularizes the meaning of a lexical item and differentiates it from the other.


1.       For details Fodor, J.D. (1977, p. 9 ff) and Southworth and Daswani (1974 p. 171 ff.) and other works on semantics.

2.       Zgusta (1971, 36.37) gives a detailed treatment of the subject.

3.       Specially the Firthian School of Linguistics.

4.       For arrangement of homonyms in a dictionary see Arrangement of Entries

5.       See Tiwari and Chaturvedi 1970.

6.       Ibid.

7.       Hornby, COD, HSS, Sanskrit Dictionary (Poona), Sanskrit-English Dictionary(Monier-Williams) etc.

8.       Examples from Ghatage (1976-78, p. XVI).

9.       The American College Encyclopaedic Dictionary gives the following 18 ways in which synonyms can be discriminated:

a. Between general and specific

b. Between shades of meaning

c.In emphasis

d. In implication

e. In application

f. In connotation

g.In emotional effect

h.In levels of usage

i. Between literacy and colloquial usage

j. Effects of prefixes and suffixes

k. In idioms

l.In British and American usage

m. Between borrowed and native words

n.Between literal and figurative usage

o.between concrete and abstractness

p. Between technical used (or occupational uses) and popular uses

q. In respect of action

r.Between local or provincial usage and general usage.  Collinson has tabulated the most typical differences between synonyms and has distinguished between nine possibilities (Ulmann 1962, Chapter 6)

10.Goodenough (1956), Lounsbury (1956), Wallace and Atkins (1960) etc.