I wrote in a previous blog post (Plausible typos and homophones) about the continued importance of human proofreaders and their superiority (for now) over computers. Another area where humans’ use of contextual inference outperforms machines’ analysis is the grammatical issue of reference – the use of pronouns and pronoun phrases to avoid repeating previously used nouns.
I remember a university lecture on machine translation that highlighted the issue with the following example:
• The monkey ate the banana because it was ripe.
• The monkey ate the banana because it was hungry.
As human speakers of English, we understand that ‘it’ refers to the banana in the first instance and to the monkey in the second, but early machine translation systems struggled to render the difference in French, for example, where ‘it’ has to be feminine in the first case and masculine in the second:
• Le singe a mangé la banane parce qu’elle était mure.
• Le singe a mangé la banane parce qu’il avait faim.
However, reference can pose problems for humans too when sentences become longer and more complicated. Take a look at the following example (with all the pronouns in bold):
• Anna steals an apple from a market stall. Her friend Bonnie, who is with her, tells her that it is wrong, but she is mistaken for the thief and is punished for the crime that she committed.
It starts off clearly enough, but at the end it is not obvious who the second ‘she’ refers to. Now try this version:
• Anna steals an apple from a market stall. Her friend Bonnie, who is with her, tells Anna that it is wrong, but Bonnie is mistaken for the thief and is punished for the crime that Anna committed.
It is now perfectly clear what happened to whom, but the repetition of the names is perhaps a bit excessive. A good solution is often to use paraphrases instead of names or pronouns to make the writing varied but keep it clear who is being referred to:
• Anna steals an apple from a market stall. Her friend Bonnie, who is with her, tells her that it is wrong, but Bonnie is mistaken for the thief and is punished for the crime that her friend committed.
Now let’s look at an example from an academic context:
• The section that follows contains an analysis of the data. Because this is based on qualitative research, it does not provide an objective, measurable picture of the issue, and they should be interpreted in the context of the whole interview with each respondent. This will enable them to get a sense of their attitudes towards the issue.
Confused? If so, it is not from a lack of subject knowledge but from the writer’s ambiguous use of reference. It is not clear whether the first ‘this’ refers back to just ‘data’, to the analysis or to the section of the paper, and, to add to the confusion, ‘data’ itself switches from singular to plural within the course of the sentence. Again, it is not clear whether the second ‘this’ refers back to the data, to the context of the whole interview or to the interpretation of that context. Finally, are ‘them’ and ‘their’ the same people or not, and who on earth are they anyway?
This version should make a lot more sense:
• The section that follows contains an analysis of the data. Because this data is based on qualitative research, it does not provide an objective, measurable picture of the issue, and should be interpreted in the context of the whole interview with each respondent. This wider context will enable researchers to get a sense of respondents’ attitudes towards the issue.
A good rule of thumb for academic writing and editing is to avoid using ‘this’ in isolation but always to specify ‘this what’. Doing so can then help untangle other problems of unclear sentence structure, and of unclear thinking behind it. My hope is that it will be many a long year before Microsoft Word can do this for you!
Share this post:
Follow us:
