Checklist and hints for preparing essays
- Don’t forget to insert page numbers on your essay
- Use the past tense to talk about what you did (not “I will use Corpus 1”). You are reporting on what you have done.
- Don’t be too narrative about what you did. Rather than “I looked at 50 tokens in corpus 2 too and found that …”, say “A similar 50-token sample of Corpus 2 showed …” i.e use a terse, concise style
- Tables and figures: don’t use charts / tables for very small raw figures (e.g a pie chart divided into three sections that is only about a total of 5 instances is a bit misleading). In tables, always include a column / row for the total figures as well as for the individual categories – helps give an overview. If giving % results, always give the raw figures as well.
- If you identify a number of categories of analysis in your essay, make sure you label these clearly in all tables and figures. Your poor reader will not remember that Category 1 means deverbal noun, 2 means derived from denominal simplex, etc….
- Remember to describe your corpus /corpora for readers who don’t already know all about them. Don’t just say “I used the Mannheimer Morgen corpus”, but “the Mannheimer Morgen corpus, a corpus of X-many words from the period Y, freely available from …..” and then go on to describe the language / content you’d expect to find in the corpus. (literature? journalism? formal/ informal, etc?). The corpora should also be listed with full details in the bibliography/ list of references.
- Don’t be afraid to give actual examples (not just the figures, but the sample words / sentences behind the figures). Analysing some examples can help you come up with an explanation for why you found what you found (e.g. in what contexts is the würde form preferred over the synthetic subjunctive?), or can illustrate any difficulties you had in trying to categorize your data
- Make sure your essay reflects the final state of your investigation – if you realize part-way through that you need to re-classify your data for some reason (e.g. you need to discount tokens of –reich that turn out not to be a suffixoid at all, but are parts of the noun Österreich), do this for all the data you present. Don’t tell us on p.9 that the table on p.2 is not right after all.