Abstract

Frequent lexical patterns can explain how language, society and culture interact. In this paper, we analyze the most frequent adjectival collocates which precede lemmas WOMAN and MAN, by searching the node words woman, women, man and men in the British National Corpus (BNC) using the statistical procedure list. The primary postulate is that frequent collocational patterns reveal common societal and cultural concepts. The research is based on Sinclair’s theory about how frequency points to what is typical and central in a language (17). Furthermore, Stubbs’s understanding of a community’s value system being built up and maintained by the recurrent use of particular phrasings in texts (Words and Phrases 166) is explored through the repetition of lexical patterns in the corpus, thus exposing dominant cultural models.

Keywords: WOMAN, MAN, BNC, frequency, collocates, language, society, culture

1. Introduction

Michael Stubbs’s principle that “language in use transmits the culture,” by which he provides his understanding of the relations between form and meaning (Text 43), is a good foundation for the study of the frequency of words by means of electronic corpora. Since meaning is language in use, electronic corpora facilitate just that – an analysis of raw, unaltered data, as clearly stated by Stubbs’s second principle concerning language being studied in “actual, attested, authentic instances of use” (Text 28).

Collocations are a fundamental organizational principle of language in use (Stubbs, Words and Phrases 60). Collocations shape important cultural concepts and provide irrefutable empirical evidence; as stated by Stubbs, “An important fact to be extracted from corpora is simply the absolute frequency of each collocation, since what we are looking for is recurrent phrases which encode culturally important concepts.” (Text 174) Therefore, a study into the frequency of words through collocations using electronic corpora can tell us about the culture of a discourse community, the perspectives and impressions of its members. Cultural values, stereotypes, ideology and inequality between social categories lie behind linguistic patterns. Aston and Burnard argue that typical collocations may appear neutral but can, in fact, hide prejudice (14). They demonstrate this by Stubbs’s analysis of high frequency collocates of Welsh and Irish, which solidify nationalistic stereotypes. Studying the word nationalist, Stubbs reasons that it is a word with negative prosody (Text 186) because it collocates with words such as demagogue, disrupt, extreme, extremists, fanaticism, etc. If nationalist is used with words which denote nationality, the negative prosody is passed on to those words. For Joanna Channell the word fat is culturally motivated, and its evaluation is culturally agreed among the members of the British cultural community (41-43). Stubbs agrees with Deborah Cameron that languages not only manifest the attitudes of society and the individual, but they also actively participate in creating them (Text 61).[1] Alan Jones also concurs: “Our language on the one hand shapes the way we perceive the world we live in and, in particular, our social world; but, at the same time, through its rich potential for creating new meanings, it allows us to act upon and shape that world.”

Stubbs’s thoughts about a community’s value system which is built up and maintained by the recurrent use of particular phrasings in texts (Words and Phrases 166) originate from John Sinclair (17). Sinclair’s views about the frequency of linguistic patterns, i.e. collocations, state that frequency portrays what is central and typical in a language. Frequent collocational patterns point to common societal and cultural concepts.

Sinclair’s and Stubbs’s work on collocations provides a foothold for linguistic analysis with the use of electronic corpora and facilitates the research of languages and what they reveal about society and culture. For instance, Michael Pearce in the text titled “Investigating the Collocational Behaviour of MAN and WOMAN in the BNC using Sketch Engine” examines the representations of MAN and WOMAN in the British National Corpus (BNC) by focusing on the collocational and grammatical behavior of the lexemes man/men and woman/women. Pearce concludes that the collocates of MAN and WOMAN represent gender in stereotypical ways, associating men with notions such as competitiveness, rationality, aggression and dominance, and women with emotional intemperance, physical weakness and subordination. Stereotypical representation of gender can be noticed in adjectives of physical appearance: men are barrel-chested, broad-shouldered or stout while women are buxom, plump or slender. In “The Company Women and Men Keep: What Collocations Can Reveal about Culture,” Sara Gesuato studies collocates of WOMAN, MAN, GIRL and BOY in four components of the Cobuild online corpus, finding that attributes which describe WOMAN and GIRL are more likely to be unfavorable or focused on the perception of beauty while those modifying their male counterparts tend to be more evenly apportioned between positive and negative qualities or emphasize men’s social roles and cerebral qualities. In Language and Society: An Introduction to Sociolinguistics, Romaine presents her research of the collocability of bachelor and spinster in the BNC, finding that the collocates of bachelor are largely descriptive or positive (106-10). The collocates of spinster are predominantly negative, with the exception of some neutral descriptive adjectives (Romaine 106-10). Likewise, her analysis of man, woman, boy and girl shows that words with negative overtones are still more frequently used together with girl and woman than with man and boy.

In reference to the aforementioned findings, the aim of this research is to contrast the nouns which denote two basic social categories ‒ lemmas WOMAN and MAN, i.e. woman and man, and women and men, by examining their collocational behavior. More specifically, the purpose is to determine whether our findings will correspond with the observations of previous research and confirm the assumption that frequent lexical patterns reveal common societal and cultural concepts.

2. Methodology

The nouns or node words were searched in the British National Corpus. In order to access it we used an interface (Davies). In order to harvest the collocates, we used the option list, which generates the hundred most frequent collocates, and the formula [j* woman] and [j* women] for the lemma WOMAN and [j* man] and [j* men] for the lemma MAN. The emphasis is on attributive adjectives and their frequency. Attributive adjectives fall into the category known as adjacent collocations, which includes collocates that occur immediately before or after the node word. Window collocations, in contrast, span the number of collocates to four or even five collocates before or after the keyword (Lindquist73-87). Many functional words are excluded through only focusing on adjectives, as advocated by Anatol Stefanowitsch and Stefan Th. Gries (1-3). They suggest that if we focus solely on adjectives, we eliminate functional words and therefore concentrate only on words that can help us understand the nouns analyzed. By examining the relationship between adjectival collocates and the nouns they modify, we can determine what nouns reveal about language and how language interacts with society and culture.

Our generated collocates were placed into semantic fields based on their meaning, and names of the fields were formulated accordingly, taking into account the research carried out by Gesuato and her classification of semantic fields. The division of adjectives into three categories ‒ positive, neutral and negative characteristics (Hlebec) – was also taken into consideration.

Bearing in mind Sinclair’s views on lemmatization (8, 41-42, 173-74), especially that “each distinct form is potentially a unique lexical unit, and that forms should only be conflated into lemmas when their environments show a certain amount and type of similarity” (8), we decided to utilize it. Liisi Piits in “Distributional Hypothesis: Words for ‘Human Being’ and Their Estonian Collocates” applies lemmatization to her corpus because Estonian contains 28 inflective suffixes for nouns. Grammatical homonymy can also pose a problem, and Piits refers to Gesuato, who also applies lemmatization. Gesuato’s conclusions about the context of woman stem from the examples of collocates of the nouns woman and women.

3. Analysis and Discussion

Table 1 Adjectival collocates of woman in the BNC

Table 2 Adjectival collocates of man in the BNC

The semantic field of physical appearance is the only category which is characterized by three subcategories: positive, neutral and negative. Neutral collocates of both lexemes include black, white, blond/e, dark, etc. Positive and negative collocates reveal what society and culture appreciate in men and women. For women, being tall and thin is appealing while fat, dumpy, plump, etc. are features that are disparaged. For men, the emphasis is on strength ‒ big, large, strong and tall. Some premodifiers are features typically reserved for men ‒ bald. Evidenced by the corpus, if women are strong, they are most likely strong in character. Therefore, their physical strength is rarely emphasized.

The field of physical attractiveness has six collocates for woman and four for man. It is interesting to note that the adjectival collocate handsome is used to describe woman and man. Man is not defined with beautiful, pretty or lovely. No collocates which comprise the negative polarity of either noun were generated. One adjective which could fall into that category is macho.

The contrastive analysis between woman and man regarding the field of age shows that the collocates are practically indistinguishable. The only collocation that stands out is 26-year-old woman. Paul Baker explains this phenomenon by saying that whatever statistical procedure is employed, once function words have been removed, atypical words seem excessively emphasized (qtd. in Pearce 4). The six remaining collocates, however, differ in frequency. Middle-aged, old, older, young, younger pattern more frequently with man while the attributive adjective elderly occurs in the corpus 125 times for woman and 108 times for man. Gesuato explains this phenomenon by invoking a taboo towards women and their age: “The euphemistic term referring to old age (elderly) is reserved to WOMAN, which signals a mild taboo in the co-occurrence of ‘femaleness’ and ‘age’.” (253)

The number, frequency and variety of collocates used to describe the character of man trumps the one of woman. Women are admired for their decency and honesty, and for their strong character. Men are as well. But there are also adjectives that appear to be reserved for men only, and they denote a sense of humor, charm and bravery. The collocates with negative meaning question woman’s morality ‒ fallen and scarlet – while men’s morality remains unchallenged. With 46 occurrences of quiet man, man is also defined as being an introvert. Rajna Dragićević argues that character qualities rank higher than physical ones (84-88), which would explain why the number of collocates in physical appearance and physical attractiveness falls behind in comparison to the field of character.

Michael Pearce employs Sketch Engine in order to investigate the collocational behavior of MAN and WOMAN in the British National Corpus. He utilizes Goldberg’s Big Five personality traits for the classification of collocates. One of the personality traits he discusses is neuroticism, whereby the author argues that women are more commonly portrayed as suffering from permanently negative mental states than men are. He argues that WOMAN patterns more strongly with distraught, dissatisfied, mad, neurotic and silly and that only WOMAN collocates with hysterical and weeping. Our approach generated the collocation mad woman appearing 27 times and silly woman appearing 17 times in the corpus. Mad man and silly man were not generated.

Pearce uses the collocation mad woman to exemplify Taylor’s view on differently distributed adjectives: “Attributive adjectives tend to characterise a thing in terms of a more stable, inherent property, whereas predicative adjectives tend to donate more temporary, circumstantial properties” (16). By contrasting mad woman with the man is mad, we see that mad woman illustrates that madness is closely connected to the noun. The man is mad suggests that his madness might only be a temporary condition or, if a complement is added, it might even have a neutral or positive connotation ‒ the man is mad about sport.

Value system, which groups together characteristics deemed favorable or unfavorable by society, is similar to character in the way it outnumbers collocates of woman in favor of man. While there are some common collocates – average, ordinary, strange and similar adjectives – men are portrayed in a rather exalted way; they are front man, head man, leading man, top man, powerful and important. Women do not have such qualities.

The collocation educated man appears 28 times in the corpus in the hundred most frequent adjectival collocates. Educated woman does not. If we search educated woman via KWIC (Key Word in Context), we get three results. If we observe the field of intelligence, positive adjectives of man are more numerous than those of woman. Stupid woman appears 18 times, stupid man does not appear. However, stupid man is also generated 18 times via KWIC. List did not produce the collocation stupid man because the last generated collocation, grey man, has 25 results.

The field of emotional state tends to describe men more in terms of their emotions in comparison with women. Men are happy, lucky or lonely while women are only unfortunate.

The semantic field love relations and marriage comprises five collocates for woman and only two collocates for man. If we look at the number of frequencies, the collocation married woman appears in the corpus 174 times, single woman 50 times, lone woman 12 times, divorced woman and unmarried woman 10 times. Man, on the other hand, is either married (158 times) or single (53 times). It is evident that women are more described in terms of their marital status than men are.

If we look at the types of the semantic fields of woman and man, they are practically identical. The only differentiating semantic field is the one that is typically female ‒ children and motherhood ‒ where women are described as pregnant or menstruating.

There are a few adjectives which are polysemic in nature and such adjectives are italicized. Poor can mean lacking money, or it can be used as an expression of sympathy. Single can refer to a person’s marital status, or it can mean individual. Collocations new woman and new man may indicate a shift in one’s personality: “Next morning I was a new woman” (B1N), or it may refer to a new, unknown person: “Roy Tolan has become obsessed with a new woman” (CH6). Different acts in the same way: “I was a different man, full of strength and excitement” (HGS), and “Tolkien was by temperament a very different man from Lewis” (A7C). A real man is either masculine, “It makes a man feel like a real man” (FU8), or real as opposed to imaginary: “and be able to tell the difference between a stuffed dummy and a real man” (AT7). The same is with woman: “but it takes a woman to hold him. A real woman” (HA9) and “yet his nervous system responds to it in the same kind of way as it might respond to a real woman” (ARR). The collocation best man refers to the man who assists a bridegroom or a man who is the best in a particular field.

The field of religion contrasts Catholic with holy and religious. Women are clearly defined in terms of their religious identity while men are either saints or they are just plain believers. Gesuato calls religious premodifiers permanent properties or attributes of appropriate conduct (254-55).

Semantic fields such as financial situation, health, death and dying are fairly similar. In terms of their number, the common premodifiers rich, poor, blind, sick, dead and dying are all in favor of men. Men’s financial situation or, as Gesuato refers to it, non-physical attractiveness (253), has two more synonyms, richest and wealthy, which suggests that society is more concerned with men being rich and that this is an important feature of manhood.

The semantic field of place of origin is significantly more detailed for women than it is for men. Men are only described as local. Pearce offers a credible explanation as to why a woman’s national identity is a matter of importance to society (12). He posits that these adjectives are not marked by gender and are therefore understood to be referring to men. The field of ethnicity is generated only for woman.

If we search for the collocation well woman via KWIC, we see that it actually refers to hospitals that concern themselves with women’s health issues. If we observe the context of invisible man, we see that the collocation refers to Ralph Ellison’s novel of the same title.

A new field concerning men defines man as human being, thus denoting both sexes, male and female, and using collocates early, primitive and wild. Due to the distinction between the two meanings of man, it is imperative to examine each concordance individually, when we have statistical procedures such as list.

The overall observation of collocates is that most of them are very generic in meaning and come as dichotomies. People are either good or bad, rich or poor, young or old. Collocates either establish the relationship of antonymy, when we look at dichotomies, or synonymy, within a polarity slot. Some adjectives, such as good and bad, can be considered prototypes. Dragićević states that such adjectives are desemantisized, central, synonymous, and that they represent language universals (84-88). She studies adjectives which denote human qualities and concludes that the centrality of adjectives is based on the appreciation and high regard of certain qualities in a discourse community. Central is what is typical in a language.

Table 3 Adjectival collocates of women in the BNC

Table 4 Adjectival collocates of men in the BNC

Physical appearance and physical attractiveness of women and men compared to their singular forms are fairly similar. Attention is again focused upon men’s physical strength while women are favored as lean rather than fat. Men is, however, missing the field of physical attractiveness. Age for women and men produces new collocates, adult, aged and senior.

The number of the premodifiers used to describe the character of women and men is lower in comparison with the singular forms of the analyzed lexemes. Men is found in positive environments in relation to intelligence: clever men occurs 12 times, wise men 115 times. If we search these collocates with women via KWIC, clever women occurs three times while wise women occurs once. Clearly, society gives greater credit to men than women as far as intellectual capacity is concerned.

The value system of men continues to depict them as key, leading, powerful and top. Women are in this instance described as leading as well, with nine results compared with leading men, which appears 26 times. Costa, et al. explain this by arguing that: “[m]en in all cultures are physically stronger than women, they may universally be assigned roles as leaders, and in these roles may learn to become more assertive than women.” (324) Regarding education, educated women is mentioned 28 times in the corpus and educated men 18 times, which implies that the education of women is an important matter of debate in social sciences, judging by the material in which this collocation appears. The emotional state of men is merry while women are not characterized by the same collocate.

Women is modified significantly often by collocates of love relations and marriage. There are nine attributes for women and only three for men. The frequency of collocates is thus as follows: married women appears 477 times, divorced women 18 times, single women 96 times, married men 70 times, divorced men 10 times, single men 37 times. These are significant findings for a corpus in which men are referred to more than women.

The biological role of women is evidenced by pregnant, primiparous and childless. Primiparous refers to women who are either pregnant for the first time or have only given birth to one child. It appears in nine concordances and is not in common use. If we look at the context, this collocate appears in medical journals and a medicine book. It would seem that the most common collocates can occasionally be misleading.

A new field introduced by women and men is sexuality. Men are more likely to be referred to with adjectives which denote homosexuality – homosexual men produces 88 results and homosexual women generates 7 results – while women are more often described with adjectives of heterosexuality ‒ heterosexual women 27 times and heterosexual men 14 times.

The collocates of financial situation, health, death and dying, place of origin and independence are fairly similar to the collocates of singular nouns. The religious identity of women is, again, a more important societal concern than the religious identity of men. The variety of adjectives concerning work demonstrates that women’s employment is an important part of women’s identity, although the two matching adjectives of employment for women and men are more frequently used with men ‒ unemployed men co-occurs 29 times and working men 71 times, compared with unemployed women (9 times) and working women (26 times).

Another differentiating field is politics, where only women are described as Conservative women, Labour women or feminist women. A possible explanation for this phenomenon is the one provided for nationality adjectives. Being Conservative or Labour is understood to be referring to a man. Feminist women does not require explaining. Conservative also means old-fashioned, as in “or what one might call the old, old fashioned conservative women” (KRL).

Since men are characterized as being physically stronger than women, which is a societal norm, women can become targets of that strength. They are depicted as being victims of violence committed by men ‒ collocates battered and raped. A semantic field that separates men from women is one with which men are usually associated ‒ military. The collocation medical men refers to men who practice medicine, with 46 results. If we search for medical women via KWIC, we get one result, “The Medical Women’s Federation has been pressing the Department of Health” (EC7).

Man as human being generated for men provides the same meaning as it did with man, as in “a pared-down Christianity among the wild men” (A6B). The plural forms of lexemes introduce another new field, class, where the collocation middle-class women has 80 results, working-class women 105 results, middle-class men 14 results and working-class men 24 results. The asymmetry between the number of collocates for women and men attests that class is an important societal construct as far as women are concerned. The adjective active is in most concordances modified by an adverb ‒ economically active, politically active or sexually active.

4. Conclusion

The analysis of the lemmas WOMAN and MAN through collocations outlined what the most frequent collocates are, and through these frequent patterns of collocations common cultural models were revealed. Physical attractiveness stands out as a semantic field more relevant for WOMAN than for MAN. Elderly is found to collocate with WOMAN only, confirming age as a taboo for female human beings, which therefore requires a euphemism.

The characters of WOMAN and MAN share some collocates. However, while the collocates of MAN are more numerous and therefore more diverse, the usage of collocates of immorality and neuroticisms is restricted to WOMAN. Society also gives more prominence to MAN, who apparently displays signs of great intellectual capacity. The fundamental distinction concerning value system is the portrayal of MAN as leading, with its synonymous collocates. Thus, the clear implication is that lexical patterns can create inequalities between WOMAN and MAN and through their repetitive usage perpetuate and strengthen them.

The gender bias toward WOMAN is reflected in love relations and marriage, where society’s perception of WOMAN is that being married is a concept upon which women are judged. Attention is given to WOMAN and her biological function. Another bias concerns sexuality. Since WOMAN is characterized more frequently with collocates of heterosexuality and MAN with those of homosexuality, a possible explanation of this tendency is that more males than females identified themselves as homosexual. Another is society’s traditional concern for the perpetuation of the species. Or, perhaps, looking at it from a historical perspective, men have always been dominating the discourse and leaving little room for women and their matters.

The number of collocates for WOMAN which form the fields of place of origin, ethnicity, politics and class exceeds those of MAN due to a lack of gender markedness. Semantic fields solely reserved for MAN are military, medicine and man as human being. Man as human being recognizes MAN as a polysemic word, cautioning us to always look at the context. The semantic field of independence consistently generates independent for WOMAN and free for MAN. It appears that the woman is more often perceived as having the resolve and confidence not to depend on external factors while men do not have to face external restrictions at all, or they have a choice not to be bound by such restrictions.

Contrary to popular expectations that women are more emotional than men, or more emotionally expressive than men, and should thus amass more adjectives of emotions, society seems to be more concerned with the emotional state of MAN, enumerating three positive adjectives and one negative. Although research exists which shows that women are indeed more emotionally expressive (Fischer and LaFrance),[2] this semantic field follows the numerical order of the corpus. According to the number of occurrences, MAN is mentioned in the corpus more than WOMAN.

How language reflects society can be observed in the collocates which represent WOMAN as a victim of violence committed by MAN. As Stubbs in Text and Corpus Analysis evokes Hall’s analysis of ideology and sees it as “the reproduction of a dominant discourse, in which particular definitions and classifications acquire, by repetition, an aura of common sense, and come to seem natural and comprehensive rather than partial and selective” (194), we can see how society and culture interact and how language serves as a medium between them. Language can be a tool used not only to retain but also strengthen those cultural models which conceal prejudice and stereotypes within a discourse community. Conversely, repetitive collocational patterns can also be seen as a reflection of cultural models existing in a society. The analysis of the hundred most frequent collocates which precede lemmas WOMAN and MAN in the British National Corpus using the method list confirms the assumption that frequent lexical patterns reveal common societal and cultural concepts. This affirms similar findings by Stubbs (Text, Words and Phrases), Pearce and Gesuato. Further research will focus on the contrastive analysis of other nouns which denote human beings, starting with the biological and societal roles of WOMAN and MAN: MOTHER, FATHER, WIFE and HUSBAND, and how their most frequent adjectival collocates compare to the collocates of WOMAN and MAN.

Works Cited

Aston, Guy, and Lou Burnard. BNC Handbook. Exploring the British National Corpus with SARA. Edinburgh University Press, 1998.

Channell, Joana. “Corpus-Based Analysis of Evaluative Lexis.” Evaluation in Text: Authorial Stance and the Construction of Discourse, edited by Susan Hunston and Geoff Thompson, Oxford University Press, 2003, pp. 39-55.

Costa, Paul T., et al. “Gender Differences in Personality Traits across Cultures: Robust and Surprising Findings.” Journal of Personality and Social Psychology, vol. 81, no. 2, 2001, pp. 322-31, doi: 10.1037/0022-3514.81.2.322.

Davies, Mark. British National Corpus (BYU-BNC). Brigham Young University, 2004, http://corpus.byu.edu/bnc/. Accessed 25 Aug. 2017.

Dragićević, Rajna. Leksikologija srpskog jezika. Zavod za udžbenike, 2010.

Fischer, Agneta, and Marianne LaFrance. “What Drives the Smile and the Tear: Why Women Are More Emotionally Expressive Than Men.” Emotion Review, vol. 7, no. 1, 2015, pp. 22-29, doi: 10.1177/1754073914544406.

Gesuato, Sara. “The Company Women and Men Keep: What Collocations Can Reveal about Culture.” Proceedings of the Corpus Linguistics 2003 Conference, edited by Dawn Archer, et al., Lancaster University Centre for Computer Corpus Research on Language, 2003, pp. 253-62.

Hlebec, Boris. “Odnos značenja, obličkih promena i komponibilnosti imenica i prideva u engleskom jeziku.” Anali Filološkog fakulteta, no. 11, 1974, pp. 431-45.

Jevrić, Tamara, and Anica Radosavljević. “Slika žene u poslovicama na engleskom i srpskom jeziku: jezik kao sredstvo promovisanja prihvatljivog ponašanja.” Srpski jezik, književnost, umetnost, edited by Dragan Bošković and Časlav Nikolić, Kragujevac Faculty of Philology and Arts, 2013, pp. 467-74.

Jones, Alan. “Michael Halliday: An Appreciation.” International House Journal of Education and Development, no. 28, 2010, http://ihjournal.com/michael-halliday-an-appreciation. Accessed 18 Aug. 2017.

Lindquist, Hans. Corpus Linguistics and the Description of English. Edinburgh University Press, 2009.

Pearce, Michael. “Investigating the Collocational Behaviour of MAN and WOMAN in the BNC using Sketch Engine.” Corpora, vol. 3, no. 1, 2008, pp. 1-29, doi: 10.3366/E174950320800004X.

Piits, Liisi. “Distributional Hypothesis: Words for ‘Human Being’ and Their Estonian Collocates.” Trames, vol. 17, no. 2, 2013, pp. 141-58, doi: 10.3176/tr.2013.2.03.

Romaine, Suzanne. Language in Society: An introduction to Sociolinguistics. 2nd ed., Oxford University Press, 2000.

Sinclair, John. Corpus, Concordance, Collocation. Oxford University Press, 1991.

Stefanowitsch, Anatol, and Stefan Th. Gries. “Corpora and Grammar.” Corpus Linguistics: An International Handbook, edited by Anke Lüdeling and Merja Kytö, De Gruyter, 2009, pp. 933-52.

Stubbs, Michael. Text and Corpus Analysis. Blackwell Publishers, 1996.

---. Words and Phrases: Corpus Studies of Lexical Semantics. Blackwell Publishers, 2002.