Het effect van lexicale taaladvisering op het Belgisch-Nederlandse taalgebruik. Een diachroon corpusonderzoek naar factoren van invloed

This dissertation, entitled The effect of lexical language norms on Belgian Dutch language use. A diachronic corpus-based study of influencing factors, is an empirical study in the domain of normative linguistics. The main research question is to what extent lexical language norms have an effect on Belgian Dutch standard language use. More specifically, it will be investigated whether there are any factors that influence the possible effect of language norms on language reality, and if so, which factors.Methodologically, this study adopts both a qualitative and a quantitative approach. The linguistic literature mentions a large number of factors that may influence the success of language advice, but their true effect has hardly been investigated empirically. For this study, eighteen factors have been selected. In the qualitative research component a large set of prescriptive reference works, language norms and lexical items has been studied over a period of fifty years (1958-2008). The results consist of the onomasiological profiles for 118 lexical nominal concepts, consistent with the concept-based approach. Because the question whether a specific word is a standard or non-standard variant was often answered more than once and sometimes also inconsistently, an artificial 'constructed norm' was created for each of the three periods in our research, viz. 1958-1978, 1978-1998 and 1998-2008. The quantitative study was designed to investigate the usage frequency of all variants in a standard context by means of a digital corpus. The corpus consists of Belgian Dutch newspaper language and allows both register stratification and diachronic stratification, containing language material from two newspapers (De Standaard en Gazet van Antwerpen) and four measuring points (1958, 1978, 1998 and 2008). The combination of these frequency data with the constructed norms from the qualitative research results in an 'error percentage' for each concept in each period, i.e. the extent to which a concept is lexicalised in a non-standard way. Finally, we analyze the error percentages in relation to the various factors that may have influenced them. By means of descriptive statistical analyses such as means, correlations and relative frequencies, we look for relations between the error percentages and the influencing factors. This study's main contribution lies precisely with this empirical approach, considering existing research on prescriptivism rarely reaches beyond the qualitative level.The research carried out generally confirms our central hypothesis: language norms have a positive effect on standard language use. From a diachronic perspective, the advantageous consequences of language norms are undeniable. Firstly, the mean error percentage continues to decrease over the years. Secondly, the number of non-standard variants in Belgian Dutch newspaper language in 2008 is only half of that number in 1958, which indicates a major influence of prescriptivism. Thirdly, the number of non-standard variants decreases to the largest extent between 1958 and 1978, precisely the period in which there was substantial social concern about Standard Dutch, in which of a long list of prescriptive works was published and in which correct language use was strongly propagated by the media. The synchronic perspective, however, warrants nuancing of the conclusion that half a century of prescriptivism has been extremely successful. Not only does the lexicalisation of the investigated concepts still show too much variation in standard and non-standard variants, the mean error percentages also remain quite high for every measuring point: they vary from around 42 % in 1958 to 21 % in 2008. However, the same synchronic data can also be interpreted more positively: on average, a concept is lexicalised by means of a standard variant in most cases.The study has shown that factors influencing the effects of prescriptivism in a positive way do exist. As we can point out factors that promote the standard realisation of concepts, we can conclude that the success of language norms is neither random nor unpredictable. The decisive factor turns out to be concept frequency. In every period language norms on concepts that frequently occur in language use are clearly applied more than language norms concerning concepts that occur only rarely in newspaper language. Put differently, the more frequent a concept occurs in language use, the higher the chance that it will be lexicalised by means of a standard variant. Concept frequency leads to variant familiarity and, as a consequence, to norm application.Besides concept frequency four more factors turn out to influence the extent of norm application quite consistently. Synchronic norm uniformity is a first influential factor. Language norms on one variant which are unanimous in content in a certain period, are applied more often than language norms on a variant which is evaluated discordantly within the same time period. Uniform advice is more successful. This holds especially for 1998 and 2008. A second important factor is the diachronic stability of language norms. In 1998 and 2008, we see that norms that have remained stable over the years are applied more often than norms that have been changed in content over time. Thirdly, register appropriateness is another influential factor. Language norms on words that are rejected as non-standard because they do not belong to a neutral register are applied more often than norms on words that are rejected because of other reasons. We can conclude that the Flemish language user has been aware of register suitability since the 1960s and 1970s.Lastly, it has become apparent that the extent of standard language use in newspapers is dependent on the time period. The language climate changes over the years, and so does the application of language advice. In general, norms are applied better as years go by: the mean error percentage in Belgian Dutch newspapers continues to decrease. The period between 1958 and 1978 proves to be beneficial to the influence of prescriptivism. This does not come as a surprise as in this period society highly values Standard Dutch and both media and prescriptive books eagerly promote the standard language. The period between 1978 and 1998 is also advantageous for the success of language norms, but to a more limited extent. Between 1998 and 2008 less non-standard variants disappear from newspaper language, although the decrease remains clearly visible. This is probably caused by the growing tolerance for Belgian Dutch in the contemporary language climate.This study also reveals five factors that have only limited influence on the success of prescriptivism: the native versus borrowed origin of words, the degree of adaptation of a loanword, stigmatization (only relevant in 1978), dialect endogeneity (only relevant in 1998) and descriptiveness (only relevant in 2008). Eight factors are not important at all, viz. the onomasiological or semasiological norm type, norm frequency, propaganda in the media, norm explicitness, norm power (the cumulation of frequency, explicitness and uniformity), perception of the source language in processes of lexical borrowing, the direct versus indirect borrowing of a loanword and newspaper specificity.In short, in the past fifty years prescriptivism has achieved important results. By taking into account the factors that positively influence the effect of language norms, prescriptive works will remain useful in the twenty-first century.nrpages: 469status: publishe