A discourse analysis of 'hope you are well?'
How to spot a fake: a discourse analysis of 'hope you are well?'
First of all - an apology
I recently - carelessly - allowed some people called University of Skills to access my LinkedIn address book. I now think that everyone I am (was?) connected to received an email from me saying something like this: 'Hey, Hope you are well ....' followed by a plug for the University of Skills website, which I wont repeat here. I apologise. I did not realise they would do that and I won't do it again.
How to spot a fake
I really do hope that the people who received this apparently personal endorsement from me are indeed well and found the website useful if they visited it. For some who received it, I am sure the endorsement looked legitimate, if somewhat surprising. That is certainly what I thought when I received it from someone else. The truth is though that it was a fake personal recommendation.
I am certain that it would have struck an odd note for some people. For example, my husband (with whom I am linked in) did wonder why I was asking after his health in such an odd fashion! People who had received an almost identical email from someone else would also have been surprised and suspicious.
There is nothing in the words 'hope that you are well' that make it seem fake, but for some people in some circumstances, fake is exactly how it would seem. In other words, whether you can spot a fake depends not on the words per se, but on the context of the interaction: the relationship between the people who are communicating, and any events that have preceded this interaction.
What can discourse analysis teach us about fakes?
Discourse analysis is the study of language in interaction. Studying discourse helps us explain what people are 'doing' with language. One of the amazing things about language, spoken or written, is about how much meaning can be packed into very small phrases, but when little things go wrong they go really wrong!
Paul Grice has shown that people have to make assumptions about each other if they are to communicate effectively. In all our communications, we first of all assume that what someone has said or written is meaningful and relevant. We only question it when there is an apparent breach of the rules. I had a boss once who said to me 'don't assume, it makes an ass out of you and me'. Clever, but wrong. We have to assume to communicate.
Once the communication seems to breach the rules, then we stop taking it at face value, and question its intent. People who I have not seen lately may have been genuinely touched by my concern for their health, which indeed I am :) However, people that I had seen that very day would have been puzzled and suspicious, thinking it was fake not because there was anything odd or abnormal about the words, but because the context for saying them was wrong.
Roman Jakobson identified 6 functions of language: referential, conative, emotive, poetic, metalingual and phatic. People use phatic expressions to open and close conversations. They are a way of saying "I am starting to speak now, pay attention." The person writing the email may have intended the phrase 'hope you are well' to be nothing more than a phatic conversational opener.
However, not all conversations need opening in exactly the same way, do they? It is not the phrase of choice for starting a conversation with my kids for example: 'Hope you are well? Please tidy your room.' seems fake, right?
Receivers of messages do not necessarily interpret them the same way the speaker or writer did. Someone who has been very ill may interpret the phrase 'hope you are well' referentially, or emotionally rather than as phatic communication.
All this makes 'hope you are well?' a nice way to re-start an existing conversation, but a very problematic way to start a mass email.
How does this relate to research?
This is blog about research after all, so I should explain how this faux pas of mine* relates to research. The greatest lesson is in how researchers judge the veracity of online conversations, when the researcher him or herself was not part of that conversation, as is the case for unmoderated social media interactions. One of the assumptions inherent in sentiment analysis for example is that the speaker's sentiment can be deduced from their words. As we have seen, that is not the case; the University of Skills tried to hide behind a facade of words, but the facade crumbled. The meaning is in the context, not the words.
My recommendation for users of sentiment analysis is to combine it with discourse analysis just as researchers would conduct small-scale qualitative research to help understand and explore large scale survey data.
.. and I do hope you are well!
*A discourse analyst would say that by confessing to a 'faux pas' (and using that term) I am using a 'pre-emptive defence' which is to use language for emotional reasons rather than convey information.
Faith in fakes
By the way, the image at the top of the blog is from a lovely book by the semiotician Umberto Eco called. One of his most famous sayings is that 'semiotics is the study of everything that can be used to tell a lie.' Think of discourse analysis as the semiotics of language.