Methods and Procedures of Lexicological Analysis

The results of the co-occurrence or distributional analysis may be of great help to teachers in preparation of teaching material.

To illustrate the point under consideration it is sufficient to discuss the experiment the goal of which was to find out the semantic peculiarities of the verb to giggle. Giggle refers to a type of laughter — to giggle is usually defined as ‘to laugh in a nervous manner’. There is nothing in the dictionary definition to indicate a very important peculiarity of the word-meaning, i.e. that giggling is habitually associated with women. A completion test carried out by a group of English linguists yielded interesting results.

The sentences to be completed were of the type: The man — with obvious pleasure, The woman — with obvious pleasure, etc. The informants were to fill in the blanks with either the verb to laugh or to giggle and were presented with a choice of subjects male and female. A clear preference was shown for women giggling and men laughing with obvious pleasure. The analysis of the informants’ responses also showed that a man may giggle drunkenly or nervously, but not happily or politely. In the case of women, however, of whom giggling is more characteristic it appears that all collocations — giggle drunkenly, nervously, happily, politely — are equally acceptable. It may be inferred from the above that the meaning by co-occurrence is an inherent part and an essential component of the word-meaning.

§ 5. Transformational Analysis

Transformational analysis in lexicological investigations may be defined as re-patterning of various distributional structures in order to discover difference or sameness of meaning of practically identical distributional patterns.

As distributional patterns are in a number of cases polysemantic, transformational procedures are of help not only in the analysis of semantic sameness / difference of the lexical units under investigation but also in the analysis of the factors that account for their polysemy.

For example, if we compare two compound words dogfight and dogcart, we shall see that the distributional pattern of stems is identical and may be represented as n+n. The meaning of these words broadly speaking is also similar as the first of the stems modifies, describes, the second and we understand these compounds as ‘a kind of fight’ and ‘a kind of cart’ respectively. The semantic relationship between the stems, however, is different and hence the lexical meaning of the words is also different. This can be shown by means of a transformational procedure which shows that a dogfight is semantically equivalent to ‘a fight between dogs’, whereas a dogcart is not ‘a cart between dogs’ but ‘a cart drawn by dogs’.Word-groups of identical distributional structure when re-patterned also show that the semantic relationship between words and consequently the meaning of word-groups may be different. For example, in the word-groups consisting of a possessive pronoun followed by a noun, e.g. his car, his failure, his arrest, his goodness, etc., the relationship between his and the following nouns is in each instant different which can be demonstrated by means of transformational procedures.

his car (pen, table, etc.) may be re-patterned into he has a car (a pen, a table, etc.) or in a more generalised form may be represented as A possesses B.

his failure (mistake, attempt, etc.) may be represented as he failed (was mistaken, attempted) or A performs В which is impossible in the case of his car (pen, table, etc.).

his arrest (imprisonment, embarrassment, etc.) may be re-patterned into he was arrested (imprisoned and embarrassed, etc.) or A is the goal of the action B.

his goodness (kindness, modesty, etc.) may be represented as he is good (kind, modest, etc.) or В is the quality of A.

It can also be inferred from the above that two phrases which are transforms of each other (e.g. his car -> he has a car; his kindness -> he is kind, etc.1) are correlated in meaning as well as in form.

Regular correspondence and interdependence of different patterns is viewed as a criterion of different or same meaning. When the direction of. conversion was discussed it was pointed out that transformational procedure may be used as one of the criteria enabling us to decide which of the two words in a conversion pair is the derived member.2

Transformational analysis may also be described as a kind of translation. If we understand by translation transference of a message by different means, we may assume that there exist at least three types of translation:3 1. interlingual translation or translation from one language into another which is what we traditionally call translation; 2. intersemiotic translation or transference of a message from one kind of semiotic system to another. For example, we know that a verbal message may be transmitted into a flag message by hoisting up the proper flags in the right sequence, and at last 3. intralingual translation which consists essentially in rewording a message within the same language — a kind of paraphrasing. Thus, e.g., the same message may be transmitted by the following his work is excellent -> his excellent work -> the excellence of his work.

The rules of transformational analysis, however, are rather strict and should not be identified with paraphrasing in the usual sense of the term. There are many restrictions both on the syntactic and the lexical level. An exhaustive discussion of these restrictions is unnecessary and impossible within the framework of the present textbook. We shall confine our brief survey to the transformational procedures commonly used in lexicological investigation. These are as follows:

1.              permutation — the re-patterning of the kernel transform on condition that the basic subordinative relationships between words and the word-stems of the lexical units are not changed. In the example discussed above the basic relationships between lexical units and the stems of the notional words are essentially the same: cf. his work is excellent -> his excellent work -> the excellence of his work -> he works excellently.

2.    replacement — the substitution of a component of the distributional structure by a member of a certain strictly defined set of lexical units, e.g. replacement of a notional verb by an auxiliary or a link verb, etc. Thus, in the two sentences having identical distributional structure He will make a bad mistake, He will make a good teacher, the verb to make can be substituted for by become or be only in the second sentence (he will become, be a good teacher) but not in the first (*he will become a bad mistake) which is a formal proof of the intuitively felt difference in the meaning of the verb to make in each of the sentences. In other words the fact of the impossibility of identical transformations of distributionally identical structures is a formal proof of the difference in their meaning.

3.    additiоn (or expansion) — may be illustrated by the application of the procedure of addition to the classification of adjectives into two groups — adjectives denoting inherent and non-inherent properties. For example, if to the two sentences John is happy (popular, etc.) and John is tall (clever, etc.) we add, say, in Moscow, we shall see that *John is tall (clever, etc.) in Moscow is utterly nonsensical, whereas John is happy (popular, etc.) in Moscow is a well-formed sentence. Evidently this may be accounted for by the difference in the meaning of adjectives denoting inherent (tall, clever, etc.) and non-inherent (happy, popular, etc.) properties.

4.    deletion — a procedure which shows whether one of the words is semantically subordinated to the other or others, i.e. whether the semantic relations between words are identical. For example, the word- group red flowers may be deleted and transformed into flowers without making the sentence nonsensical. Cf.: I love red flowers, I love flowers, whereas I hate red tape cannot be transformed into I hate tape or I hate red.1

Transformational procedures may be of use in practical classroom teaching as they bring to light the so-called sentence paradigm or to be more exact different ways in which the same message may be worded in modern English.

It is argued, e.g., that certain paired sentences, one containing a verb and one containing an adjective, are understood in the same way, e.g. sentence pairs where there is form similarity between the verb and the adjective.

Cf.: I desire that. . . — I am desirous that . . .; John hopes that . . . — John is hopeful that . . .; His stories amuse me . . . — are amusing to me; Cigarettes harm people — are harmful to people.

Such sentence pairs occur regularly in modern English, are used interchangeably in many cases and should be taught as two equally possible variants.

It is also argued that certain paired sentences, one containing a verb and one a deverbal noun, are also a common occurrence in Modern English. Cf., e.g., I like jazz — > my liking for jazz; John considers Mary’s feelings -> John’s consideration of Mary’s feelings.2

Learning a foreign language one must memorise as a rule several commonly used structures with similar meaning. These structures make up what can be described as a paradigm of the sentence just as a set of forms (e.g. go — went — gone, etc.) makes up a word paradigm. Thus, the sentence of the type John likes his wife to eat well makes up part of the sentence paradigm which may be represented as follows John likes his wife to eat well — > John likes his wife eating well — > what John likes is his wife eating well, etc. as any sentence of this type may be re-patterned in the same way.

Transformational procedures are also used as will be shown below in componental analysis of lexical units.

§ 6. Componental Analysis

In recent years problems of semasiology have come to the fore in the research work of linguists of different schools of thought and a number of attempts have been made to find efficient procedures for the analysis and interpretation of meaning.3 An important step forward was taken in 1950’s with the development of componental analysis. In this analysis linguists proceed from the assumption that the smallest units of meaning are sememes (or semes) and that sememes and lexemes (or lexical items) are usually not in one-to-one but in one-to-many correspondence. For example, in the lexical item woman several components of meaning or sememes may be singled out and namely ‘human’, ‘female’, ‘adult’. This one-to-many correspondence may be represented as follows.

The analysis of the word girl would also yield the sememes ‘human’ and ‘female’, but instead of the sememe ‘adult’ we shall find the sememe ‘young’ distinguishing the meaning of the word woman from that of girl. The comparison of the results of the componental analysis of the words boy and girl would also show the difference just in one component, i..e. the sememe denoting ‘male’ and ‘female’ respectively.

It should be pointed out that componental analysis deals with individual meanings. Different meanings of polysemantic words have different componental structure. For example, the comparison of two meanings of the noun boy (1. a male child up to the age of 17 or 18 and 2. a male servant (any age) esp. in African and Asian countries) reveals that though both of them contain the semantic components ‘human’ and ‘male’ the component ‘young’ which is part of one meaning is not to be found in the other. As a rule when we discuss the analysis of word-meaning we imply the basic meaning of the word under consideration.

In its classical form componental analysis was applied to the so-called closed subsystems of vocabulary, mostly only to kinship and colour terms. The analysis as a rule was formalised only as far as the symbolic representation of meaning components is concerned. Thus, e.g. in the analysis of kinship terms, the component denoting sex may be represented by A — male, A — female, В may stand for one generation above ego, В — for the generation below ego, С — for direct lineality, С — for indirect lineality, etc. Accordingly the clusters of symbols ABC and ABC represent the semantic components of the word mother, and father respectively.

In its more elaborate form componental analysis also proceeds from the assumption that word-meaning is not an unanalysable whole but can be decomposed into elementary semantic components. It is assumed, however, that these basic semantic elements which might be called semantic features can be classified into several subtypes thus ultimately constituting a highly structured system. In other words it is assumed that any item can be described in terms of categories arranged in a hierarchical way; that is a subsequent category is a subcategory of the previous category.

The most inclusive categories are parts of speech — the major word classes are nouns, verbs, adjectives, adverbs. All members of a major class share a distinguishing semantic feature and involve a certain type of semantic information. More revealing names for such features might be “thingness” or “substantiality” for nouns, “quality” for adjectives, and so on.

All other semantic features may be classified into semantic markers — semantic features which are present also in the lexical meaning of other words and distinguishers — semantic features which are individual, i.e. which do not recur in the lexical meaning of other words. Thus, the distinction between markers and distinguishers is that markers refer to features which the item has in common with other items, distinguishers refer to what differentiates an item from other items. The componental analysis of the word, e.g., spinster runs: noun, count-noun, human, adult, female, who has never married. Noun of course is the part of speech, meaning the most inclusive category; count-noun is a marker, it represents a subclass within nouns and refers to the semantic feature which the word spinster has in common with all other countable nouns (boy, table, flower, idea, etc.) but which distinguishes it from all uncountable nouns, e.g. salt, bread, water, etc; human is also a marker which refers the word spinster to a subcategory of countable nouns, i.e. to nouns denoting human beings; adult is another marker pointing at a specific subdivision of human beings into adults & young or not grown up. The word spinster possesses still another marker — female — which it shares with such words as woman, widow, mother, etc., and which represents a subclass of adult females. At last comes the distinguisher who has never married which differentiates the meaning of the word from other words which have all other common semantic features. Thus, the componental analysis may be represented as a hierarchical structure with several subcategories each of which stands in relation of subordination to the preceding subclass of semantic features.

This may be represented in the graphic form as

Componental analysis with the help of markers and distinguishers may be used in the analysis of hyponymic groups.1 In the semantic analysis of such groups we find that they constitute a series with an increasingly larger range of inclusion. For example, bear, mammal, animal represent three successive markers in which bear is subordinated to mammal and mammal to animal. As one ascends the hierarchical structure the terms generally become fewer and the domains — larger, i.e. the shift is from greater specificity to greater generic character. Words

that belong to the same step in the hierarchical ladder are of the same degree of specificity and have all of them at least one marker — one component of meaning in common. They constitute a series where the relationship between the members is essentially identical.

Componental analysis is also used in the investigation of the semantic structure of synonyms. There is always a certain component of meaning which makes one member of the synonymic set different from any other member of the same set. Thus, though brave, courageous, fearless, audacious, etc. are all of them traditionally cited as making up a set of synonymic words, each member of the set has a component of meaning not to be found in any other member of this set. In a number of cases this semantic component may be hard to define, nevertheless intuitively it is felt by all native speakers. For instance, that is how the difference in the meaning components of the words like, enjoy, appreciate, etc. is described. Analysing the difficulty of finding an adequate translation for John appreciates classical music; he doesn't appreciate rock the author argues that “... appreciate is not quite the same as enjoy or like or admire or take an interest in though quite1 a number of semantic components making up their meaning is identical. To appreciate is to be attuned to the real virtue X is presupposed to have and not to appreciate is to fail to be attuned. It is not to deny that X has virtues. In short, appreciate seems to presuppose in the object qualities deserving admiration in a way that like, admire, and so on do not."

Componental analysis is currently combined with other linguistic procedures used for the investigation of meaning. For example, contrastive analysis supplemented by componental analysis yields very good results as one can clearly see the lack of one-to-one correspondence not only between the semantic structure of correlated words (the number and types of meaning) but also the difference in the seemingly identical and correlated meanings of contrasted words.

For example, the correlated meanings of the Russian word толстый and the English words thick, stout, buxom though they all denote broadly speaking the same property (of great or specified depth between opposite surfaces) are not semantically identical because the Russian word толстый is used to describe both humans and objects indiscriminately (cf., толстая женщина, (книга), the English adjective thick does not contain the semantic component human. Conversely stout in this meaning does not contain the component object (cf. a thick book but a stout man). The English adjective buxom possesses in addition to human the sex component, and namely, female which is not to be found in either the English stout or in the Russian толстый. It can be inferred from the above that this analysis into the components animate / inanimate, human male / female reveals the difference in the comparable meanings of correlated words of two different languages — Russian and English — and also the difference in the meaning of synonyms within the English language.

The procedure of componental analysis is also combined with the semantic analysis through collocability or co-occurrence as the components of the lexical (or the grammatical) meaning may be singled out by the co-occurrence analysis. It is assumed that certain words may co-occur in a sentence, others may not. The co-occurrence of one word with another may be treated as a clue to the criterial feature of the concept denoted by the word. Thus, for example, if one learns that a puffin flies, one can assume that a puffin is animate and is probably a bird or an insect.

A close inspection of words with which the prepositions occur brings out the components of their meaning. Thus, e.g., down the stairs is admitted *down the day is not; during the day is admitted but *during the stairs is not. We may infer that time feature is to be found in the preposition during but not in the meaning of down. We can also see that some prepositions share the features of space and time because of their regular co-occurrence with the nouns denoting space and time, e.g. in the city / country, in July / in 1975, etc.

A completion test in which the subjects have a free choice of verb to complete the sentences show that, though in the dictionary definitions of a number of verbs one cannot find any explicit indication of constraints, which would point at the semantic component, e. g. animate — inanimate, human — nonhuman, etc., the co-occurrence of the verbs with certain types of nouns, functioning as subjects, can be viewed as a reliable criterion of such components. For example, in the sentences of the type The cows — through the fields, The boys — through the fields, etc. various verbs were offered stray, wander, ran, lumber, walk, hurry, stroll, etc. The responses of the subjects showed, however, the difference in the components of the verb-meanings. For example, for all of them stroll is constrained to human subjects though no dictionaries include this component (of human beings) in the definition of the verb.

The semantic peculiarities of the subcategories within nouns are revealed in their specific co-occurrence. For example, the combination of nouns with different pronouns specifies the sex of the living being denoted by the noun. Cf. The baby drank his bottle and The baby drank her bottle where the sex-component of the word-meaning can be observed through the co-occurrence of the noun baby with the possessive pronouns his or her.

Componental analysis may be also arrived at through transformational procedures. It is assumed that sameness / difference of transforms is indicative of sameness / difference in the componental structure of the lexical unit. The example commonly analysed is the difference in the transforms of the structurally identical lexical units, e.g. puppydog, bulldog, lapdog, etc. The difference in the semantic relationship between the stems of the compounds and hence the difference in the component of the word-meaning is demonstrated by the impossibility of the same type of transforms for all these words. Thus, a puppydog may be transformed into ‘a dog (which) is a puppy’, bull-dog, however, is not ‘a dog which is a bull’, neither is a lapdog ‘a dog which is a lap’. A bulldog may be transformed into ‘a bulllike dog’, or ‘a dog which looks like a bull’, but a lapdog is not ‘a dog like a lap’, etc.

Generally speaking one may assume that practically all classifications of lexical units implicitly presuppose the application of the the-ory of semantic components. For instance the classification of nouns into animate — inanimate, human — nonhuman proceeds from the assumption that there is a common semantic component found in such words as, e.g., man, boy, girl, etc., whereas this semantic component is nonexistent in other words, e.g. table, chair, pen, etc., or dog, cat, horse, etc.

Thematic classification of vocabulary units for teaching purposes is in fact also based on componental analysis.

Thus, e.g., we can observe the common semantic component in the lexico-semantic group entitled ‘food-stuffs’ and made up of such words as sugar, pepper, salt, bread, etc., or the common semantic component ‘non-human living being’ in cat, lion, dog, tiger, etc.

§ 7. Method of Semantic Differential

All the methods of semantic analysis discussed above are aimed mainly or exclusively at the investigation of the denotational component of the lexical meaning.

The analysis of the differences of the connotational meaning is very hard since the nuances are often slight, difficult to grasp and do not yield themselves to objective investigation and verification.

An attempt to establish and display these differences was developed by a group of American psycholinguists.1 They set up a technique known as the semantic differential by means of which, as they claim, meaning can be measured. It is perfectly clear, however, that what semantic differential measures is not word-meaning in any of accepted senses of the term but the connotational component of meaning or to be more exact the emotive charge.

Their technique requires the subjects to judge a series of concepts with respect to a set of bipolar (antonymic) adjective scales. For example, a concept like horse is to be rated as to the degree to which it is good or bad, fast or slow, strong or weak, etc.

The meaning of the seven divisions is, taking as an example the first of the scales represented above, from left to right: extremely good, quite good, slightly good, neither good nor bad (or equally good and bad) slightly bad, quite bad, extremely bad.

In the diagram above horse is described as neither good nor bad, extremely fast, quite strong, slightly hard, equally happy and sad.