, ,

by Arnold Wentzel


Having previously argued that the literature review is not a ‘review’, but rather a series of connected arguments in support of the research question, I now explain how to generate the arguments that need to be connected to each other and to the literature. The person who asks a research question makes a number of assumptions, and finding the assumptions that underlie the question generates the raw content of a literature review. By arguing that each assumption is plausible, and linking them logically to each other, one adds substance to the literature review.

Finding assumptions in any question is easier said than done, even for experts. What is required is called shoshin in Zen Buddhism or ‘beginner’s mind’. In such a state there is an awareness of preconceptions and an ability to be detached from them. The act of thinking about a literature review encourages one to become aware of the taken-for-granted ideas that can imprison one’s thinking, go beyond it and treat it as being open to argument.

One way to learn how to find assumptions is to scaffold the process with a rather mechanical approach I call ‘relational analysis’ that helps to analyse the relations between the terms in an operational research question.

Before this technique can be used, the question needs to be operationalised. This means the question should be defined so well that it meets three conditions: (1) the question by itself suggests what information is required to answer the question; (2) the question by itself suggests where and when to get that information; and (3) the question by itself suggests by what method to obtain or generate the information. One way to do this is to take each noun, verb, adjective and adverb in the question and define what is meant by it until the question meets the three conditions. After such a process a question like ‘what is the secret to success?’ may become ‘how can an Australian today accumulate $10m in a bank account by age 35?’

The next step is to break the question up into its key terms. The question may be separated into four terms: (1) method (the how); (2) Australia; (3) $10 million in the bank; and (4) age 35. There is some discretion here, but a researcher familiar with the literature should know which terms are most relevant. Since all the terms appear in the same question, the questioner assumes that there is a relation between them. This is then mapped as a relation diagram in the figure below.

Arnold 1

The more key terms, the more relations, so it usually more manageable to work with only three to five key terms. Each of the pair-wise relations (labelled A to F) is then questioned to determine what is assumed about the relationship between each pair of terms in the question. In the case of relation A one may ask: ‘what is the relation between method and $10 million that is assumed in the question?’ and realise it assumes that there is some method that one can follow to obtain financial success.

Once a list of assumptions has been generated, it is necessary to identify the mundane or obvious assumptions. These assumptions will either not be argued or will be discussed as background in arguing the other assumptions and can be removed from the list. Any assumptions that are very similar should also be joined at this point.

The remaining assumptions, which I label ‘key assumptions’, are assumptions about which there are different perspectives. The question usually favours one perspective over others, and hence each such assumption will call for an argument to be made.

Doing this for each relation, one might emerge with the assumptions below (after combining duplicate assumptions and removing those that are too obvious):

A: There is a method that guarantees financial success

B: Any method takes time to deliver results

C: Most Australians will live much longer than 35

D: Australians evaluate success in terms of money

E: Success is only meaningful if you are still young

F: Australia is predictable enough for a method to work

The list above shows only six assumptions, but it is quite possible for more than one assumption to be generated for each relation, or that some relations may not have any non-obvious assumptions. A researcher familiar with the literature is more likely to produce a meaningful list of assumptions, as this step involves judging which assumptions are arguable, and often finding assumptions not directly evident from the relation diagram.

This is also be the best time to determine if any of the assumptions are implausible. If so, the research question needs to be changed because a single implausible assumption makes it impossible to find a meaningful answer to the question.


In the next step the researcher will have to argue that each of the non-obvious assumptions is plausible with respect to the literature the audience accepts. Each assumption will therefore be treated as the claim of an argument.  At this point you will have generated only the claims (assumptions); now you need a writing plan that tells you how to connect them logically and elaborate on these claims.