"Drunk" or "drunken". A sociolinguistic research on two adjectives

Table of contents

1. Introduction

2. Previous research/theoretical background

3. Data and methodology
3.1 Data
3.2 Methodology

4. Results and analysis
4.1 Results and analysis of social patterns
4.2 Results and analysis of linguistic patterns
4.2.1 Quantitative results
4.2.2 Qualitative analysis

5. Discussion

6. Conclusion

References

1. Introduction

This paper sets out to examine the use of the adjectives drunk and drunken. According to the Oxford English Dictionary the definition of drunk is “Affected by alcohol to the extent of losing control of one’s faculties or behaviour” (OED, s.v. drunk, adj.), and that of drunken reads “Drunk or intoxicated” (OED, s.v. drunken, adj.). Consequently, their use on the lexical level seems to be the same and, therefore, interchangeable. Both expressions are used in temporary English and are not antiquated as a search in the BNC (=British National Corpus) shows (BNC 2018). Thus, the question may arise whether other factors determine the application of each of these adjectives. The field of sociolinguistics deals with such questions of variation in language or more precisely in the use of different expression denotating the same meaning, while it does not simply assess different utterances as either correct or incorrect (Tagliamonte 2012: 2). Sociolinguists like Sali A. Tagliamonte (2012: Contents) rather examine language influencing variables like discourse features, social patterns (gender, ethnicity etc.) or linguistic patterns (syntax, morphology etc.), of which the two latter ones will serve as the investigative frame of the two adjectives drunk and drunken. In order to probe thoroughly into the usage of these two words, this paper, firstly, delivers a look into previous research on, and into the theoretical background of the respective adjectives. Secondly, the original data, namely the British National Corpus, and the gained random samples from this corpus are characterized, whereafter the methodology of coding for social (accent, sex and social grade) and linguistic (various syntactic functions) features is explained. The results and analysis of this coding for factors which may influence the usage of drunk and drunken are, thirdly, presented quantitatively and qualitatively in the two sections of social patterns and of linguistic patterns. These two analyses are discussed separately as well as in view of the respective other patterns to gain the deepest possible insight into the findings of this linguistic research. Finally, this paper is completed by a conclusion, which rounds off the linguistic intention to examine whether social and linguistic variables influence the usage or preference of the synonymous adjectives drunk and drunken.

2. Previous research/theoretical background

The variable use of same or similar adjectives in different syntactical or social contexts has been studied from various (socio-)linguistic angles. Numerous social studies focus at least on one of the three social variables which are relevant to the following analysis. Especially the field of comparative sociolinguistics has been exploring dialectical differences in either the usage of a certain word or in the expressing of a certain meaning (Tagliamonte 2012: 162). One of these researches has been carried out by Hughes, Trudgill and Watt (2013), who show that there are significant dialect variations within the borders of the British Isles like the positions of direct and indirect objects, where in northern England more liberty in the order of these is possible (Hughes, Trudgill and Watt 2013: 20).

Concerning the social pattern of biological sex, various studies such as the one by Beth Thomas, who investigated differences in speech amongst a Welsh mining village (Cameron and Coates 2014: 51-60), result in the conclusion that women are more sensitive to differentiated language use than men, especially when it comes to socially stigmatised forms (Labov 1972: 243). Several explanation models may be used to interpret such results; one explanation, according to Chambers (2003: 149-153), may be that women possess a superior biological ability for linguistic features compared to men.

One of the most prominent social patterns which will also be of relevance in this paper is social class, also called social grade. The National Readership Survey (NRS) developed a still valid and useful social classification for Great Britain which consists of the grades A (higher managerial, administrative and professional = upper middle class), B (intermediate managerial, administrative and professional = middle middle class), C1 (supervisory, clerical and junior managerial, administrative and professional = lower middle class), C2 (skilled manual workers = upper working class), D (semi-skilled and unskilled workers = lower working class) and E (state pensioners, casual and lowest grade workers, unemployed with state benefits only) (NRS 2018). That these social classification system is not only valid for population censuses but also for linguist research show the manifold comparative studies on differing language use between social classes (Tagliamonte 2012: 25-32).

In contrast to the field of social investigations, the linguistic, and more precisely the syntactical perspective has already looked upon drunk and drunken. However, it seems that the respective researches do only touch on this concrete subject rather than focus on it. In his conference paper on English participle-adjective conversion, Joan Bresnan (1995: 8) mentions the origin of the adjective drunken as an older participle form in a mere footnote. Teyssier (1968: 238) refers to the two adjectives in his linguistic article in a more detailed manner and claims that drunken “remain[s] purely classifying and non-predicative” and that drunk, consequently, is only used in an attributive meaning without supporting this claim with numerical or other evidence. Biber, Conrad and Leech (2002: 188) define an attributive adjective as preceding and modifying the head noun as part of the noun phrase, whereas a predicative adjective does not belong to a noun phrase but follows a copular verb (most commonly be) or occurs as an object predicative. Another considerable syntactic role of adjectives are the free modifiers which modify a noun phrase but are not a part of it and are mostly placed at the beginning of a sentence (Biber, Conrad and Leech 2002: 203). Whether an adjective can only be used predicatively or attributively or in both ways determines whether it is classed as ‘peripheral’ (only one of the suggested uses is possible) or ‘central’ (attributive and predicative use is conventional). Typical examples for peripheral adjectives are utter, which is exclusively used attributively, and alive, which is only fit for predicative use (Biber, Conrad and Leech 2002: 189; Matthews 2014: 2-3). Free modifiers do not appear in common linguistic discussions on the classification between ‘peripheral’ and ‘central’.

When it comes to modifying adjectives, it seems logical to consider the modified noun as well. The general classification of main types of noun reads ‘abstract noun’ (abstraction like states, times and qualities), ‘proper noun’ (singular and definitive like names and specific places), ‘collective noun’ (group of people, items or animals) and ‘common noun’ (entities which are general and not abstract) (Biber, Conrad and Leech 2002: 56-61). According to the studies of Biber, Conrad and Leech (2002: 199), certain trends occur with the use of attributive adjectives, which mainly modify common nouns, less commonly proper place nouns and hardly ever proper nouns which are the names of persons. The respective tendencies of predicative adjectives are untried or at least not mentioned in the pertinent works.

This short overview of previous research and grammatical theories illustrates certain gaps, open questions and a lack of numerical evidence, which will be pursued in the following analysis of the use of the two adjectives drunk and drunken.

3. Data and methodology

3.1 Data

The data for this study originates from the British National Corpus (BNC), namely from the latest version of 2007 named ‘BNC XML Edition’. This corpus is limited to modern British English from the late twentieth century and consists ninety percent of written language from different newspapers, specialist and fiction books, school essays or letters, for instance. The remaining ten percent of data is spoken and transcripted language which was collected from informal and formal conversations as well as radio shows. As this corpus encompasses data from all styles, registers, ages and the like, it is classified as ‘general’ (BNC 2018).

In order to ensure that all searched for tokens of the types drunk and drunken are only adjectives and not past participles, the tag ‘_JJ’ is added to the input of the BNC’s search engine. That said search process with tagging is not wholly flawless, as shows a swift scanning of the first hundred results, where isolated past participles appear. For the analysis of the social variables, which are already filtered by the BNC, this minor defectiveness remains a reservation but is not be solved as the whole BNC data for each of the adjectives is used. That is to say that the verification and reassessment of 321 tokens for drunken and 1645 tokens for drunk does not lie within the scope of this paper and, thus, this faultiness is accepted but borne in mind.

Since the analysis of the linguistic variables which influence the use of drunk and drunken bases on a distinctly developed coding scheme and is, therefore, more laborious, the sample sizes are reduced to 100 tokens for each adjective. For this analysis’ faultlessness sake, a random sample of 150 words of each type is gathered, exempted from all past participles and finally cut down to 100 tokens, which are then used for the coding. Such reduced random sample sizes certainly entail the results not being utterly representative of these adjectives’ general uses; nevertheless, this research tries to give an idea of their average usages.

3.2 Methodology

As for the analysis of social parameters, the already coded classification of the BNC is used for all of the three considered patterns (dialects, sex, social grade). Dialects or accents are encoded and listed as can be seen in Table 1.

Abbildung in dieser Leseprobe nicht enthalten

The social pattern of sex is subdivided into the categories ‘F’ (female), ‘M’ (male), ‘U’ (unknown) and ‘Not_Set’ (no sex information entry for this word). Finally, the social grades are coded approximately according to the NRS classification but less distinguished. Thus, the social grade coding in the British National Corpus reads ‘AB’, ‘C1’, ‘C2’ and ‘DE’, as well as ‘UU’ (unknown) and ‘Not_Set’. For the analysis of these social features and their distribution amongst the tokens of drunk and drunken, the filtering option ‘Frequency Distribution’ of the BNC is used.

For the investigation of linguistic variables, the two random samples for the two types drunk and drunken are copied into an Excel file, where they are, as mentioned in section 3.1, checked on truly being adjectives. In order to decide on the respective word classes, a form of world class test, which is used by many syntacticians and asks, among other things, for the possible modification with very, is applicated (Vartiainen 2016: 63-65). The entries of two sample sizes are then coded into two main categories, both of which follow the classifications of Biber, Conrad and Leech (2002: 188, 203, 56-61) as listed in section 2; firstly, which syntactical function they fulfill (predicative use, attributive use or free modifier); secondly, which kind of noun they modify or, especially with predicative instances, what the respective subject is (abstract, proper, collective, common, gerund or a pronoun, of which the latter two noun or pronoun types are added for the sake of completeness). By means of examples from the two

[...]