Introduction
Attitudes to English Usage is the title of a book published in 1970 by W. H. Mittins, Mary Salu, Mary Edminson and Sheila Coyne from the University of Newcastle upon Tyne that reported on an enquiry held among some 450 informants concerning the acceptability of 55 usage items. These items had been selected because they were at the time ‘subject to variation in practice and dispute in theory’ (Mittins et al., Reference Mittins, Salu, Edminson and Coyne1970: 4), and they include sentences like He refused to even think about it, It looked like it will rain and Everyone has their off-days. In each case the offensive feature had been highlighted so that informants would know what they had to comment on: to even think (a split infinitive), the use of like for as if, and of their with a singular antecedent (everyone). For fifty sentences the informants had to indicate acceptability in informal speech, informal writing, formal speech and formal writing, and for the remaining five only for informal and formal writing, since usage of these items was believed to be restricted to writing (1970: 4). The sentences were subsequently ranged from highest general acceptability (did not do as well as) to lowest (veryunique), and correlations were calculated with the occupation of the informants (students, teachers, lecturers, examiners and non-educationists), while the items were also classified as colloquial (pretty reliable), etymological (data is), grammatical (did it quicker) and lexical/semantic (inferred/implied), or as language myths, ‘where the censorious tend to invoke a prescription of dubious authority’ (dangling participles) (1970: 15). The main part of the book dealt with the individual constructions, analysing the reasons for their status as debated usage items and providing further historical context in the process.
By the authors’ own admission, there were a number of shortcomings to the survey. One of these was the restriction of the final five items to a written context (1970: 4), while more significant reservations concerned the size of some of the respondent groups (particularly those of salesmen/advertisers/public relations officers and of professional writers) and the fact that time had been too short to carry out a more detailed correlation of the informants’ attitudes with age (1970: 21). There are, however, other shortcomings of the survey. One is that the factor of gender was not taken into account: in any sociolinguistic study, gender is a standard correlate to try and account for variation in usage (see e.g. Chapter 7 in Mesthrie et al., Reference Mesthrie, Swann, Deumert and Leap2009), so the possibility should have been considered that men and women would have responded differently. Another important issue is the representativeness of the informants: the majority (87%) comprised people involved with education: teacher-trainee students, teachers, lecturers and examiners (1970:18). The small group of non-educationists consisted of informants working in industry and commerce, public relations and administration, writers, and people from a non-educational sector. It would not be unreasonable to assume that educationists would have responded differently to questions relating to linguistic issues than informants with other professional concerns.
Developing a different approach
The study by Mittins et al. did show an extremely high response of the informants originally approached: of the more than five hundred people invited to participate, well over ninety per cent did so (1970: 5). There was clearly considerable interest in the topic of the survey, both among educationists and non-educationists. The Mittins survey, however, was conducted well over forty years ago, and since then we have learnt much about the risks of undertaking direct questionnaires like this one, as informants rarely express their true, unbiased opinions on the use of a particular feature, especially when they are aware of its disputed status. In the Mittins survey the features were, moreover, highlighted, which left no doubt as to the focus of the question addressed. In the early 1970s, Labov identified the existence of the notion ‘overt prestige’, which ‘refers to positive or negative assessments of variants ... in accordance with the dominant norms of the public media, educational institutions and upper middle-class speech’ (Mesthrie et al., Reference Mesthrie, Swann, Deumert and Leap2009: 89). Particularly when asked to express their attitudes to (highlighted) features of disputed usage, as in the case of the Mittins survey but also in that of Sandred (Reference Sandred1983), informants’ responses would be biased against features which they knew, however dimly, to clash with accepted standard practice.
A new approach to the study of attitudes to usage is therefore needed, and in this paper I report on an experiment carried out in 2012 which sought to elicit responses from informants from different backgrounds on the acceptability of a small number of usage items. Informants were asked to produce short pieces of texts that could be analysed electronically for such things as the expression of positive or negative attitudes to the item in question. The experiment took place in the context of the research project ‘Bridging the Unbridgeable: Linguists, Prescriptivists and the General Public’ (http://bridgingtheunbridgeable.com/), led by me and currently in progress at the Leiden University Centre for Linguistics. The project studies the age-old and persistent clash of attitudes towards prescriptive and descriptive approaches to language held by the three target groups of its title. To give an example, searching the electronic news media database Factiva for the occurrence of what is a virtual icon of prescriptivism, the split infinitive, I came across the following comment:
The split infinitive used in the penultimate paragraph by the editorial board of the New York Times is so disappointing I could barely finish reading this piece (New York Times Online, 29 January 2010; emphasis added, here as well as throughout this paper).
The split infinitive may be illustrated with the sentence ‘To boldly go where no man had gone before’ from the trailer of the popular series Star Trek, but also with the message that appears when users exit Microsoft Office Outlook: ‘Are you sure you want to permanently delete all the items ... ?’ Criticism of its use is drawn from the notion in Latin grammar that an infinitive is a single grammatical unit that should not be split (Mittins et al., Reference Mittins, Salu, Edminson and Coyne1970: 72). The notion clashes with the fact that in English an infinitive consists of two words, to and the verb proper (go, delete), which throughout the history of English have regularly appeared with constituents like adverbs (boldly, permanently) in between the two elements (Mittins et al., Reference Mittins, Salu, Edminson and Coyne1970:70). It is only since the 1830s that that usage began to evoke criticism, first of all in the New England Magazine (Beal, Reference Beal2004: 112). The iconic nature of the split infinitive is evident from the fact that it is often drawn upon to illustrate the notion of prescriptivism: The Language Wars (Hitchings, Reference Hitchings2011), for instance, opens with a chapter called ‘To boldly go’.
The reader commenting on the split infinitive in the above quotation from the New York Times is probably not a linguist: most linguists would adopt the perspective that usage needs to be described rather than be affected by prescriptions such as the stricture against the split infinitive, even if usage goes against prescribed norms of correctness. The reader may be what in our research project we term a prescriptivist, a teacher, a text writer or editor – in short, anyone professionally concerned with correct usage. More likely, however, he or she belongs to the category we have labelled ‘General Public’, which basically comprises anyone except linguists and prescriptivists, people who write so-called ‘letters to the editor’ to complain about usage when notions of grammatical correctness are violated. This practice is a well-established phenomenon in English culture (Cameron, Reference Cameron1995: viii), which is characterised by what Milroy and Milroy (Reference Milroy and Milroy1999) call the ‘complaint tradition’. Such a tradition, these authors argue, focuses on ‘relatively trivial’ language issues, and is typically found ‘in technologically advanced societies which require a heavily codified standard language’. English speakers, according to Hitchings (Reference Hitchings2011: 4), are indeed ‘touchy about questions of usage’, and English could even be ‘the most contested language’ in this respect.
The attitudes survey was published online, as a post on the Bridging the Unbridgeable blog (7 May 2012). Informants were also invited to participate directly, particularly by contacting colleagues, friends and acquaintances (native and non-native speakers of English), which subsequently snowballed through second-order contacts. One particularly helpful offer for distributing the request came from the former director of ELT publishing at Cambridge University Press, and founding father of English Today, Adrian du Plessis, who placed a call on the website of the University of the Third Age (http://www.u3a.org.uk/). This call was highly successful, and ensured that the online survey did not solely attract younger informants. Another very helpful offer came from internet guru Jeff Jarvis, author of What Would Google Do? (Reference Jarvis2009), to distribute the request among the members of his social network; this action accounts for an interesting group of ICT and related specialists among the respondents.Footnote 2
The survey
For the survey I selected three sentences:
1. I could of gone to that party.
2. Their errors will likely be in their use of style words.
3. He only had one chapter to finish.
The web form contained the following instructions:
We are interested in what you think about this sentence. Is it acceptable in English today, would you use it yourself? If so, where and when? If not, why not? If you think the sentence is unacceptable, why would that be the case? Do you ever hear (or see) people using it? What kind of people? Do you object to anyone using it?
Please tell us about all this in a short piece of text in the box below, which we will be able to use in our research about attitudes to usage. Thank you!
Selecting the first sentence had been inspired by a discussion with a British colleague in linguistics whose sixteen-year-old daughter had been highly surprised to learn that of in (1) was not a preposition but an auxiliary verb. The sentence in (2) is a quotation from Pennebaker (Reference Pennebaker2011: 29), on which book more below; in private correspondence from the mid-1990s between two linguistic colleagues the use of likely had been raised as a potential new usage issue that was criticised as possibly an Americanism. In contrast to the practice in Mittins et al. (Reference Mittins, Salu, Edminson and Coyne1970), likely had not been highlighted, so many informants did not identify the feature, commenting on the use of style words instead. The placement of only, the issue illustrated by (3), represents an ‘old chestnut’, a regular feature in usage guides (cf. Weiner, Reference Weiner, Stanley and Hoad1988: 173; Peters, Reference Peters, Aarts and McMahon2006: 760). It is Item 21 in the Mittins survey, and it had a general acceptability rate of 45% at the time.
The online survey was accompanied by blog polls for all three sentences in which readers, along the lines of the Mittins survey, could indicate their preferences. Multiple preferences were allowed so the total number of informants expressing their opinions cannot be verified. Around the end of July 2013 the results for the placement of only looked as presented in Figure 1.
Figure 1. Blog poll on the acceptability of He only had one chapter to finish.
Though acceptability for the different styles is not uniform, no informant marked the sentence as ‘unacceptable under any circumstances’, which suggests that acceptability of the construction must be considerably higher than in the 1970s. The results for the blog poll on likely may be found in Figure 2 and those for could of in Figure 3.
Figure 2. Blog poll on the acceptability of Their errors will likely be in their use of style words.
Figure 3. Blog poll on the acceptability of I could of gone to that party.
The fairly even spread of the votes in Figure 2 suggests that the use of likely is not considered problematical, even though a fair proportion of the informants thinks the usage unacceptable. In contrast to my colleague's impression alluded to above, Burchfield, in his third edition of Fowler's Modern English Usage (Reference Burchfield1996), so likewise in the mid-1990s, did not perceive it as a usage problem either, merely observing that ‘in standard [British?] English [adverbial likely] is almost always qualified by another adv., esp. more, most, quite, or very but just as often stands without an adverbial prop in AmE’ (Burchfield, Reference Burchfield1996: 460). Opinions therefore do not appear to have changed much over the past twenty years with respect to this feature. The figures in Figure 3, however, indicate that could of (instead of could have) is perceived to be problematical: 18 of the 27 votes condemned it as ‘unacceptable under any circumstances’. The blog poll survey thus identified a new usage problem, one that to my knowledge does not yet feature in any usage guide.
In the survey proper, informants were invited to write a short text about the acceptability of the sentences concerned. In adopting this line of approach I was inspired by Pennebaker (Reference Pennebaker2011). James Pennebaker is a social psychologist who developed a method – as well as the electronic tool Language Inquiry and Word Count (LIWC) – by which he could analyse people's disorders and general states of minds in so far as they can be assessed through their language use. He was thus able to demonstrate that during the previous American presidential elections one candidate's use of articles suggested that he was a more dynamic thinker while the other, by contrast, was shown to be more categorical in this respect (Pennebaker, Reference Pennebaker2011: 297–9). Features that Pennebaker takes into account for his linguistic assessments are such things as the use of self-references (I, me, mine), social words (mate, daughter, neighbour), words for positive or negative emotions (love, hate, kill, sad), overall cognitive words (know, think, guess) and so-called ‘big’ words, words longer than six letters. Reactions to usage problems tend to evoke considerable emotional response – inducing readers to write letters to the editor, for instance – and this is something I decided to focus on in analysing the survey results. The following comment, from a 65-year-old British female psychotherapist on the sentence I could of gone to that party, presents a fairly typical response: ‘Unacceptable, because “of” replaces “have” and is just a kind of grammatical sloppiness or laziness rather than creativity with language.’ A similar assessment occurs in the following passage from an introductory textbook on sociolinguistics, which discusses the concept of prescriptivism:
Prescriptivism, the dominant ideology in language education, holds that changes in language norms occur to the detriment of the language, and are a result of sloppiness, laziness and a lack of attention to logic. Sociolinguists feel that there is thus a need for a more scholarly understanding of the processes of change and their social contexts (Mesthrie et al., Reference Mesthrie, Swann, Deumert and Leap2009:110).Footnote 3
Inattention to logic is a complaint that also came up in the responses to the questionnaire:
It is illogical and ungrammatical and indicates poor understanding of the language. It is not acceptable and I would not use it myself. I hear people using this very often, and at my workplace I used to see it in writing from others quite frequently. The kind of people who make this error are possibly those who do not read very much (71-year-old male British retired flight operations engineer; could of).
This response, moreover, includes a social comment: the alleged error is attributed to ‘those who do not read very much’.
The response
Between the launch of the survey in May 2012 and early October that year, when I started to analyse the data obtained, 642 forms had been returned. Of these, the majority (337) had been submitted for could of, 171 for only and 134 for likely. These differences confirm that the informants felt most strongly about could of gone compared to standard could have gone, that they probably identified the placement of only as a stock usage item and that many failed to see likely as a potential usage problem. The texts received ranged from one-word assessments (e.g. ‘Fine!’ on the placement of only, 55-year-old female American editor, and ‘acceptable’ on could of, 68-year-old female British linguist) to the following one, on only (93 words):
As rendered, he ‘possessed’ no other demands on his time whatsoever. If the ‘only’ is moved to precede ‘one chapter’, the ‘only’ restricts his task in relation to the book, rather than overall. It is a frequent error and sloppy, but so widespread that it might seem simply pedantic if someone who used it were corrected. Those who have learned English as a Foreign Language seem to have a far better understanding and command of how the language works than those who are born into it. This may be allied to the present fashion for many people – young or politically/artistically ambitious – to aspire not to ape their ‘betters’, as in the past, but to be ‘acceptable’ to the masses. Viz. Blair's glottal stops when speaking to those he considered were familiar users themselves, and Cameron's removal of his tie in his first political appearances as Tory leader. One can only presume the motive was to fool the electorate that they ‘speak their language’ or that ‘we are all in this together’. ‘Sheep’ will simply follow their lead and the rest of us will become ever more cynical and despairing (76-year-old female British teacher).
This long text illustrates how a response to a linguistic issue may lead to other criticism, in this case the language of politicians.
Two-thirds of the informants were female (427/642), while the age range varied from 16 to 91, with a peak for people in their fifties and sixties (Figure 4). The large number of informants aged between 51 and 70 is partly due to the publication of the survey request on the website of the University of the Third Age, but the informants in this age category are far from all British. Language criticism is known to increase with age. This is evident, for instance, in the number of linguistic complaints the BBC receives, usually from Radio 4 listeners, who tend to come from the South of England and are generally ‘older and more middle-class’ (Luscombe, Reference Luscombe2012: 170). When people reach middle age, they look back on their earlier days, when, they believe, writers still knew how to spell correctly, didn't violate the rules of grammar, and things were generally better. In relation to the present-day situation, Baron (Reference Baron2002) writes that ‘we are raising a generation of language users ... who genuinely don't care about a whole range of “language rules”’. This, she argues, has resulted in what she calls the ‘whatever’ generation. Similar observations were found in the attitudes survey as well. I thus believe that the data obtained do not necessarily skew the representativeness of the survey: after all, the younger generation have a greater interest in the social media (the survey was initially published as a blog post). Targeting older potential informants as done here may well have balanced the results in a relatively realistic way. For all that, it will be impossible to draw any conclusions with respect to potential representativeness of the data obtained.
Figure 4. The informants classified by age.
In addition to gender and age, the informants were asked to indicate whether they were native speakers of English, and if not, which linguistic model they followed, British, American or ‘other’. Of the total number of informants (642), 500 were native speakers, 327 British, 130 American and 43 ‘other’ (e.g. ‘I am an English and Drama high school teacher in Australia’). As for the non-native speaker informants (142), 96 professed to follow a British linguistic model, 19 American and 27 another one. The informants were also invited to specify their profession. The reason for this was the basic research question underlying the Bridging the Unbridgeable project: to study attitudes to correctness in language use among linguists, prescriptivists and the general public. The categories presented were linguist (‘linguists’); editor, copywriter, translator, teacher and writer (‘prescriptivists’); and other (‘general public’). The informants categorised themselves as follows: linguist (74), editor (33), copywriter (7), translator (71), teacher (134), writer (25) and other (298). When classified according to the target groups, it turned out that linguists (11.5%) were significantly under-represented: 42% of the informants (270) belong to the category Prescriptivists and 46% to that of the General Public (298).Footnote 4 That linguists should be under-represented was to be expected: they are the kind of users that strongly oppose the notion of prescription. Cameron (Reference Cameron1995: 5), for instance, notes that many introductions to general linguistics begin by taking an explicit stand in the prescription–description debate (see also Bolinger, Reference Bolinger1980: 134). One of the challenges of the Bridging the Unbridgeable project is precisely to overcome this imbalance in the interest in questions relating to prescriptivism. The professional category ‘other’ is of particular interest due to the enormous variety of members from the general public who participated in the survey:
retired (unspecified), student (English, French; PhD), software developer, scientist (engineer, statistician), town planner, personal trainer, psychotherapist, policy adviser, software engineer, computer programmer, technology executive, military, wine consultant, web creative, researcher, physicist, evaluator, marketing, IT consultant, chemist, nurse, army soldier, shiftless historian, architect, accountant, mathematician, secretary, property manager, physician, coordinator of studies, civil servant, lawyer, attorney, pediatrician, foreign officer, British Council, bookseller, stage door keeper, diplomat, geophysicist, psychologist, police officer, business consultant, food service manager, PA, US army soldier, archaeologist, journalist, public health research, tourism management, case manager, pharmacist, flight operations, zoologist, Zen Buddhist teacher ...
Analysing the data
To analyse the texts, I used WordSmith Tools. This concordancing program has three main functions: it compiles frequency and alphabetical lists, studies words in their context (concordance searches), and analyses the so-called keyness of words in a text by comparing it to a reference corpus. Keyness analyses bring to light what a text is about more objectively than conventional analysis, which usually focuses on so-called content words (Culpeper, Reference Culpeper2009). Such analyses may demonstrate that even grammar words can be ‘key’ in a text. To be able to identify keywords in a text, a (larger) reference corpus is needed with which the text to be analysed is compared.
Compiling a frequency list of the collected survey texts (30,850 words) showed that I was most frequent, followed by it, the, to, of, in, is, would, and and not (all grammar words). Compared to the list of most common words in English in Wikipedia, this list is unusual. In the Wikipedia list, which is based on the one billion words in the Oxford English Corpus,Footnote 5 the ten most frequent words are the, be, to, of, and, a, in, that, have and I (all grammar words). That I should be the most frequent word in the attitudes survey texts demonstrates the personal nature of the texts submitted. Content words occur in much higher positions in the attitudes survey texts than in most frequency analyses, with the words sentence, acceptable and English appearing in positions 18, 19 and 20. Acceptable (294 instances), however, collocates most frequently with not (81 times), as in This is not acceptable English (54-year-old female editor; variety ‘other’) and It is not acceptable in writing (57-year-old female British teacher), so it does not always have a positive meaning. English (284) collocates most frequently with spoken (50) and written (40) as the first word on the left: I think it badly spoken English (61-year-old female non-native trilingual secretary; British) and This may not be strictly correct in written English (72-year-old male retired publisher; variety ‘other’). While the last quotation is phrased neutrally (not ... strictly correct), the one cited before that reflects a strong opinion (badly). Strong attitudes to usage occurred frequently in the survey, as in the following quotation on the sentence with could of:
A truly horrible example of chav-speak at its worst. This execrable abomination is (iirc...) usually accompanied by ‘your’ as an abbreviation for ‘you are’. I used to think that the greengrocer's apostrophe (‘potato's’) was as bad as it gets, but I'm starting to sink into despair at the future of the English language. ......Innit? (M 54 Br accountant).
Such strongly negative evaluations suggest what has come to be referred to as ‘moral panic’, a theory developed by the sociologist Stanley Cohen (1942–2013) towards the end of the 1960s (McEnery, Reference McEnery and Archer2009: 95). Moral panic is described by McEnery as ‘an alarmist debate around [episodes in the media and society] that … leads to action being taken to resolve the perceived problem’. The use of specific lexical items is part of the theoretical model that McEnery has developed, in particular what he calls moral panic rhetoric: ‘negatively loaded modifiers such as “filthy”, “revolting”, “brutal”, “irresponsible”, “weak”, and “degradation” being used to amplify the objects of offence’ (McEnery, Reference McEnery and Archer2009: 96). Such negatively loaded modifiers, highlighted in the above quotation, are present in many of the survey texts as well.
McEnery analysed the presence of moral panic rhetoric in the writings of Mary Whitehouse (1910–2001), a British schoolteacher who campaigned against perceived moral degeneracy in society, which she blamed on sex, violence and the media (Oxford Dictionary of National Biography). McEnery compiled a corpus based on her writings, the Mary Whitehouse Corpus (216,289 words), and carried out a keyword analysis of the texts using the Lancaster-Oslo-Bergen corpus as a reference corpus. His analysis produced high-frequency keywords like bbc, sex, television, broadcasting, sexual, programmes, programme, pornography, children, public, violence and TV, which illustrated Mary Whitehouse's alleged causes for Britain's degenerate state. Since the survey texts I had collected contained similar moral panic language to that which McEnery had identified in the Mary Whitehouse corpus, I was curious to see whether my informants would likewise blame particular elements in society for the linguistic degeneracy many of them commented on. To enable me to test this, McEnery kindly lent me his corpus for use as a reference corpus. Not surprisingly, given the topic my informants had written about, my keyword analysis resulted in many linguistic words (highlighted in bold below) that were key in the survey texts but that were not present in the Mary Whitehouse Corpus; language, after all, was not an issue in Mary Whitehouse's texts.
I, use, it, sentence, would, acceptable, english, correct, object, written, spoken, not, gone, probably, sounds, have, using, myself, think, used, chapter, could, this, likely, could've, don't, it's, finish, usage, hear, is, writing, only, unacceptable, incorrect, party, grammar, say, style, speech, language, write, wouldn't, errors, people, form, grammatically, verb, educated, but, formal, uneducated, heard, meaning, word, error, native, common, informal, speaker, ’ve, construction, grammatical, be, mistake, read, fine, like, should, never, sound, spelling, because, instead, yes, ok, context, words, although, wrong, however, definitively, pronunciation, might, find, anyone, speakers, often, sloppy, I'm, seems, or, contexts, do, lazy, text, adverb, dialect, contraction, poorly, prefer, sentences, someone, version, I'd, perfectly, younger, if, auxiliary, infinitive, adjective, tend, maybe, proposition, depending, internet, correcting, conversation, pronounced, awkward, non, american, hearing, guess, mostly, though, dutch, phonetic, pedantic, person, slightly, frequently, imagine, difference, sure, class, I've, shortened, correctly, ambiguity, ambiguous, abbreviated, incorrectly, illiterate, colloquial, phrase, can't, speaking, makes, chapters, clumsy, print, texts, tense, indistinguishable, replace, noticed, slang, mis, writer, students, saying, misspelling, understandable, unclear, rather (166 keywords)
In addition, the list contains keywords that are typical of the type of texts analysed and not of the moral issues Mary Whitehouse wrote about: acceptable, unacceptable, incorrect, fine, wrong, sloppy, lazy, poorly, prefer, correcting, awkward, pedantic, correctly, incorrectly, clumsy and unclear. Two of these, sloppy and lazy, were indeed used by Mesthrie et al. (Reference Mesthrie, Swann, Deumert and Leap2009) in their description of prescriptivism (see above). The most striking keywords in this list in relation to the question as to what forces my informants might have held responsible for the poor linguistic standards they described are educated, uneducated, class and illiterate. The prevalence of these words in the survey texts thus suggests that, contrary to the moral panic rhetoric used by Mary Whitehouse, the discussion of attitudes to correctness in language is linked to social class and education or the lack of it. Three examples of comments will illustrate this:
Acceptable. Would not naturally use it myself, as the use of ‘likely’ here sounds American to me. I would use ‘probably’ in the same position. No class implications as far as I know, but may be informal (?) (63-year-old male British copywriter.)
The sentence is not acceptable but is widely used, though I used to hear it more in South Africa than here. It arises because of the abbreviated form of ‘could have’ being ‘could've’, which sounds quite like ‘could of’ and is taken by the ignorant to be what they are trying to say. I would not object out loud to anyone using it but I would note their lack of education and file them as of a lower class than me. (65-year-old male British bookseller)
This sentence betrays the ‘speaker's’ lack of understanding of their own language's grammar. They write as they speak and cannot distinguish between ‘have’ and ‘of’. In their spoken English these words sound the same. As in English the spoken language can be taken as an indicator of class and education, I would guess the speaker would be lower class and at secondary level maximum as far as education is concerned. (66-year-old female British teacher.)
There are fifteen instances of the word class in the survey texts, all produced by British informants, which means that a link between correct language use and class is very likely a British phenomenon. Cameron (Reference Cameron1995: 93) indeed notes that ‘[i]t is frequently assumed that grammar, at least in Britain,’ – by which is usually meant standard grammar – ‘is essentially a symbol of class’. She also comments on the ‘exaggerated respect’ usually felt for standard grammar among the middle classes in the UK (1995: 107). Some of the informants in the survey lay the blame for what is perceived as a decline in standards of linguistic correctness on the changes which the educational system underwent during the 1970s and 80s (see also Cameron, Reference Cameron1995: 93):
I don't know whether this study will be researching educational trends, but I'd argue that scrapping the systematic teaching of grammar in UK state schools in the 1970s was the beginning of the end, and there's no way we can ever recover from that... my generation onward have essentially learned English by ear, and in my view the sample sentence in question is a prime example of that (35 F Br teacher).
Apart from the keywords mentioned, which were obtained through a comparison with the Mary Whitehouse Corpus, the survey texts contain many other words that express strong emotions, and the following list presents a selection of them:
amazed, angry, annoyed, careless, clumsy, clunky, condemn, confused, contamination, corrupt, crazy, creeping, degenerated, demeaning, difficult, dreadful, glaring, gobbledegook, grates, grit, hard, harshly, hideous, hurts, illogical, impossible, incredibly, infuriated, irritate, jarring, meaningless, misheard, mismatches, misplaced, mispronounced, misuse, nonsense, nuts …
Cameron (Reference Cameron1995: 120) alludes to the fact that language criticism such as that which I am reporting on here has its roots in the eighteenth century, and it is indeed striking how the terms expressing negative attitudes to usage are often identical to those employed by eighteenth-century grammarians. Comparing my data with a selected number of comments collected from eighteenth-century grammars by Sundby et al. (Reference Sundby, Bjørge and Haugland1991) illustrates this (words found in both lists have been highlighted):
bad English, careless, childish, confused, corrupt, creeping, disapproved, embarrassed, erroneous(ly), error, familiar, glaring, grate, harsh(ly), imprecise, improper, inaccurate, incorrect, mistake, nonsense, not correct, not good English, obscure, odd(ity), old, old-fashioned, offensive, unacceptable, unclear, wrong, awkward …
More analysis will have to show the precise extent of the overlap, and will also allow us to construct a special prescriptivist dictionary for the program LIWC with which we will be able to assess the prescriptive nature of specific texts. Further analysis will also be carried out in order to correlate the observed moral panic lexis with the different groups of informants, primarily to compare the linguistic views of language professionals (linguists, editors, translators and the like) with those held by the general public, and also to see if women do indeed evaluate questions of linguistic correctness differently from men.
Conclusion
This paper reported on an experiment conducted in the context of the research project ‘Bridging the Unbridgeable: Linguists, Prescriptivists and the General Public’ as an attempt to develop a new method to elicit attitudes to questions of usage. In addition to finding a different approach from that of Mittins et al. (Reference Mittins, Salu, Edminson and Coyne1970), it was also an experiment in exploring the Web as a means for eliciting data for analysis. This, I believe, proved successful, though thanks to the efforts of individuals who actively helped produce the snowball effect I needed. As a standalone survey it would probably have failed to produce sufficient data.
The discussion presented here represents only the start of a large-scale analysis with which I hope to be able to illustrate different attitudes to usage questions by the different groups the project is focusing on. Cameron (Reference Cameron1995: 85) notes that the kind of moral panic that I studied in this paper ‘address[es] symptoms rather than causes’. It is, however, the intention of the Bridging the Unbridgeable project to try to discover what causes moral panic in language by studying its very symptoms, by identifying and analysing these symptoms as well as people's attitudes to them. This paper, finally, inaugurates a regular new feature in English Today, which will invite readers of all backgrounds – linguists, prescriptivists, the general public – to contribute to our work by sharing with us their attitudes and views on an assortment of usage issues. We are very much looking forward to future input from readers.
INGRID TIEKEN-BOON VAN OSTADE has a chair in English Sociohistorical Linguistics at the Leiden University Centre for Linguistics. She has published widely on Late Modern English and on the English standardisation process, particularly on the codification of the language of the period. She is the director of the NWO-funded research project Bridging the Unbridgeable: Linguists, Prescriptivists and the General Public. Her book on the language of Jane Austen's letters will be published by OUP in February 2014. Email: I.M.Tieken@hum.leidenuniv.nl
