The current study explores language use in selected Afrikaans poems from a corpus linguistics perspective. The aim is to determine whether the poems can be distinguished from other general Afrikaans texts, other Afrikaans literature texts and from one another (poems written by men versus poems written by women). By analysing the corpora from a corpus linguistics perspective it is possible to determine whether the poets make use of specific grammatical or lexical characteristics that are unique to the corpora as a whole, but also to the male and female poets. Thus, language use and gender are investigated within the genre of Afrikaans poetry.
This study is based on previous research concerning gender and differences in language use as well as gender and genre. Differences in language use with reference to gender and genre have been researched from a variety of perspectives (Mulac 1975; Mulac and Lundell 1994; Barrette 2004 and Herring and Paolillo 2006). However, only a few studies focus on the Afrikaans language and specifically Afrikaans poetry. Poetry is an interesting genre to research from a linguistic perspective since it conceptualises language differently from other text genres. In poetry, language is probably used in the most creative way and words are often chosen very carefully. This study illustrates that certain characteristics can be identified that not only separate the two main corpora from other genres, but also separate the poetry of male and female poets. This suggests that there are both grammatical and lexical characteristics that could be attributed to male and female poets.
The current study focuses on two corpora of selected Afrikaans poems written from 2002 to 2010. Both corpora consist of 10 datasets (volumes of poetry) by male and female poets. These two corpora (the main corpora) are compared with a corpus of general Afrikaans texts, a corpus of Afrikaans literature texts, another corpus of Afrikaans poetry and also to each other. The corpus analysis program Antconc is used to analyse the different comparisons using keyword lists, concordance lines and concordance plots.
The keyword list analysis compares a target corpus with a reference corpus to determine whether there are words in the target corpus that are significantly key compared with the reference corpus. Significant keyness is indicated through a log-likelihood ratio where a statistical value of 3,84 or higher indicates a p-value of p < 0,05. Any words listed as significantly key were analysed to determine their context and whether they are distributed throughout the corpora. The context of a word is determined by generating concordance lines, while the distribution is determined through the concordance plot tool. This tool shows the number of datasets in which a word occurs. For the purposes of this study, a word has to occur in more than 50% of the datasets in order to be considered distributed throughout the corpora. Only words that are distributed throughout the corpora are considered as possibly characteristic of specific corpora. By using the concordance lines to determine the context of the words it is possible to determine whether the words are grammatical or lexical and to divide the lexical words into certain categories. These categories are considered the possible lexical categories of the two main corpora in each comparison. The grammatical characteristics are considered separately.
The comparison of the main corpora and the corpus of general Afrikaans texts indicates a significant use of certain pronouns in the main corpora. Conjunctions are more frequently used in the main corpora as compared with the general Afrikaans texts. 39 lexical categories are distinguished when the main corpora and the general Afrikaans texts are compared. These include words relating to the body and body parts, the senses, cognition, longing, etc.
When the main corpora and the corpus of other Afrikaans literature texts are compared a similar pattern emerges concerning the grammatical characteristics. However, the log-likelihood ratios are lower, which suggests that there is a smaller difference between the main corpora and the other literature texts. This is to be expected, since the corpus of Afrikaans literature texts consists of dramas, novels and prose. The comparison between the main corpora and the corpus of literature texts indicates a few words that can be added to the lexical categories since they are evenly distributed, but they fit into already existing categories mentioned above. This second comparison produces 34 possible lexical categories compared with the 39 categories in the previous comparison. No new categories are identified, but some of the categories in the previous comparison are no longer represented. There are also minor changes to the words included in certain categories. This finding suggests that certain grammatical and lexical characteristics of the main corpora remain constant, even when they are compared with different corpora of various genres.
A third comparison between the main corpora and another corpus of Afrikaans poetry supports the statements above and proves that there are certain characteristics that are shared within a single genre. The keyness is lowest when the main corpora and the additional poetry corpus are compared, suggesting that the poets use similar words in both instances. However, the grammatical characteristics of the main corpora are more or less the same as in the previous comparisons. The lexical characteristics are slightly different with small changes to the categories. This suggests that the poets in the main corpora use specific words in specific ways that allows them to be distinguished from a variety of other corpora, including another poetry corpus.
To determine whether gender plays a role in language use differences, the main corpora are compared with each other. The comparisons indicated that male poets in this corpus tend to use more pronouns, especially pronouns referring to others. It also appears that the male poets use the negative nie more often. In terms of the lexical characteristics seven categories are identified in the male/female comparison and 13 in the female/male comparison. Although some categories correspond, there are differences. The poems by male poets contain the categories action, religious concepts and death, which are not as prominent in the poetry of female poets. The female poets can be distinguished through the inclusion of the following categories: cognition, creative processes, food, house or dwelling, nature, colours, fire and light.
Based on the results, the conclusion is drawn that the differences in language use identified are determined by both genre and gender. Although the characteristics that are identified in this study do not reflect the characteristics of all Afrikaans poetry as a whole or all Afrikaans poetry by male or female poets, it does offer insights into possible characteristics of Afrikaans poetry by male and female poets.
Keywords: Afrikaans poetry; corpus linguistics; female poets; gender differences; language use; male poets
Lees die volledige artikel in Afrikaans: Die grammatikale en leksikale eienskappe van geselekteerde Afrikaanse gedigte