a resource for studying literary theory

Literary Linguistics – Analysis

In literary linguistics, different types of methodologies, that is ways of approaching a text, can be used or even combined with each other when analysing literary texts. In this analysis, I will make use of a corpus stylistic methodology. That is to say that I will use software developed for corpus linguistics in a study of literary linguistics; in particular I will use the corpus tools available through WebCorp. WebCorp was specifically designed to study language on the world wide web as it offers a large amount of texts which contain real language in use that is most up-to-date. However, in addition to this original aim of WebCorp, it can also be used very easily and efficiently for stylistic analyses, as I will show below.

The main benefit of WebCorp is obviously that it is available for free on the internet and can be used by anyone, anywhere, anytime. In order to do a stylistic analysis on WebCorp, one only needs the text to be studied in electronic form, accessible through a URL.

WebCorp - Wordlist for Goblin Market
Now, the first type of analysis one can do on WebCorp to reach initial insights into a literary text, like a poem, is a wordlist. This can be done by specifying the webpage on which the text appears or by inserting the text to analyse through copy and paste; an advantage of the latter is that any other words visible on a webpage, like for example in the menu or in ads on the side, will not be included in the wordlist. WebCorp then creates a list of all the words appearing in the chosen text, in this case ‘Goblin Market’, and indicates how frequently they occur. For the poem ‘Goblin Market’ the five most frequent words are Laura (27), buy (26), come (25), Lizzie (24), and goblin (18). This immediately identifies the main characters and the topic of the poem: the sisters Laura and Lizzie encounter the goblins and seem to be involved in an activity that revolves around coming and buying.

WebCorp - 4-grams for Goblin Market
In a next step, I decided to look at the two verbs come and buy in more detail. I used the Ngram function for wordlists on WebCorp, which provides an output of word clusters comprising any number of words between 1 and 5, i.e. Ngrams, and found that the combination come buy is indeed the most frequent two-word cluster in the poem with 21 occurrences. Even when the span of words is expanded to 3 or 4, the most frequent clusters comprise the verbs come and buy, which reveals that they are repeatedly used in the construction come buy come buy. The imperative form of the verbs points to their use in direct speech, which is to say that they are being addressed to someone, in this case potential buyers like Laura and Lizzie. This is also reflected in the fact that the most frequent 5-word Ngram is cry come buy come buy, which indicates that these words are used by the market criers, the goblins.

WebCorp - 3-grams for Goblin Market
When exploring the Ngrams further, in addition to the come and buy combination, there are other frequent clusters that reveal aspects about the setting and the characters in the poem, like the clusters along/down the glen and the goblin men. Furthermore, the clusters should not and must not signal that the sisters are aware of the dangers that surround them but in the end do not listen to their own advice.

The wordlist and Ngram analysis in WebCorp allows the user to include or filter out stopwords, like articles or pronouns. When including stopwords in the wordlist, one can note that the word like appears very frequently in the poem ‘Goblin Market’ and taking a closer look at the constructions in which it occurs, one can see that it is mainly used to introduce comparisons, as in ‘one crawl’d like a snail’. Thus, it can be said that many descriptions in the poem follow the same pattern of ‘x is like y’. Furthermore, when looking at the structure of the poem, it turns out that these like constructions cluster in some of the stanzas with like and or like introducing new lines repeatedly (see, for example, stanzas 4 and 26).

WebCorp - Search for 'Laura' in Goblin Market
In addition to the wordlist function, the collocation or search function of WebCorp can be used to find out more about the individual characters. This is done by searching for the name of a specific character, like Laura or Lizzie, and restricting the site under advanced options to the text studied (leave out http:// if it forms part of the URL). The output one gets is a list of all attestations of a character’s name in context, the default setting being a span of 50 characters to the right and left. For the poem ‘Goblin Market’, the collocations for Laura show that she is depicted as a character wanting to find out more about the goblins. Thus, she is described as “[c]urious Laura [who] chose to linger”, who “bow’d her head to hear”, “rear’d her glossy head”, “stretch’d her gleaming neck”, “[p]ricking up her golden head”.

WebCorp - Search for 'Lizzie' in Goblin Market
Lizzie, on the other hand, is described as “tender” and the opposite of her sister as she “veil’d her blushes”, “cover’d up her eyes”, and “could not bear [t]o watch”; she warned her sister (“You should not look at goblin men”), vehemently refused to look at the goblins (“‘No,’ said Lizzie, ‘No, no, no’”) and urged Laura to move away from the goblins. Thus, just by looking at the collocations of the characters’ first names, that is by studying the words with which their names co-occur, one can arrive at first insights into their characterization.

Overall, the above brief analysis of the poem ‘Goblin Market’ has shown that corpus stylistic methods can be used to approach a text from different perspectives in order to reach insights into, for example, the theme, characters, setting, actions and events which make up a literary text.