I tend to be sceptical of analyses of the linguistic form in psychiatry. It too often leads to stigmatising conclusions that, for example, patients with certain diagnoses speak ‘ungrammatically’ (I have written about in this post). So, when I saw another article which probes the linguistic form of auditory verbal hallucinations which appeared in my Twitter timeline, I approached it with some caution.

I am disappointed to say that the article “The linguistic signature of hallucinated voice talk in schizophrenia” doesn’t avoid the judgements as to how what patients said compared with the expectations of ‘normal conversational speech situation’, as they put it.

As a linguist, I am not clear at all what a ‘normal conversational speech situation’ might be. For example, is it a pub talk, family argument or perhaps confessions of love in an intimate situation? And if so, I am unclear why a patient reporting what they’re hearing would not be classified as ‘normal conversational speech situation’. In any case, I am yet to read a paper in, shall I say, ‘psychiatric linguistics’ that doesn’t make assumptions that ‘we’ speak in this wonderful grammatically correct way in which everything is beautifully connected and logical, why ‘they’ do the opposite, which needs to be uncovered for inspection and further othering of people in distress.

I could and (perhaps should) end the blogpost here, but as most of the page is still empty, I thought I’d offer some further critique of the study. Mostly, because I would like to suggest that papers such as these should be reviewed by linguists.

So, my first issue is the very basis of the paper. I don’t think (like not at all!!) that the paper investigates auditory verbal hallucinations (AVHs). Yes, the authors say they do, but they actually don’t.

Before I tell you why, I suggest that you do an exercise. Do two recordings. One of, say, a full long news bulletin and the second of you repeating verbatim of what you hear, as the bulletin goes on. I bet my bottom dollar that as you go along there will be many discrepancies. Sometimes more, sometimes fewer, but as you go along you will distort what has been said.  This is because repetition verbatim is difficult and, indeed, this is why, for example, simultaneous translation is called interpreting – it’s interpretation that actually happens.

With this in mind, the idea that patients are simply able to reproduce what they hear verbatim, with no distortion, is untenable, especially that it is also likely that what they hear is significant and emotionally charged for them. Moreover, some of what the informants hear might be abusive, vulgar, insensitive – and I would like to suggest that normal social graces don’t disappear on the shrink’s or researcher’s say-so. This is not like reproducing BBC news…. It is also likely that some will do it better than others. Moreover, the assumption that the informants tell researchers everything they hear is just nonsense and it goes right against basic methodological assumptions of any interview research. People are not books that researchers open and read.

Of course, I fully understand that you might disagree with me. In fact, I might well be wrong. My response to it is – what’s you evidence and the onus on you, and the authors? There is no access to what exactly the informants hear; I might be wrong, I might be right, we’ll actually never know. What we get is what the informants say, no less, but also no more. And it is really quite irritating to suggest that while, obviously, we are talking about what ‘they’ say, we also endow ‘them’ with the wondrous ability to represent the world without distortions.

And here is the crux of the matter. The paper is not about the linguistic features of AVHs, but about informants’ accounts of them. That still, presumably, makes the paper interesting, but the interest lies somewhere else. The paper doesn’t analysis AVHs! Moreover, it can’t do, and, at least at the moment, no paper can. I do wish researchers in psychiatric linguistics got that.

For space and my laziness, I am not going to do a full review of the methodology. Could I just suggest that ‘personal content’ (whatever it is) is not a linguistic category, just as ‘impersonal content’ is not? Moreover, the example “Spain is a democracy” suggests quite a lot of ambivalence as to how the authors understand the word ‘impersonal’, which in linguistics has a very different meaning from ‘sentences stating facts….’.

But I would like to offer some comments on the authors’ conclusions. What I find quite surprising is that the author really don’t stop to consider the context in which their informants say things. The authors’ consideration of the grammatical person and its occurrence suggests that they simply assume they have access to the real thing and no editing on the part of the informants is possible.

Once again, I think such assumptions are highly implausible and I would suggest that the absence of grammatical connectivity may well result from self-censorship. In other words, informants report what they feel comfortable with reporting and don’t give ‘grammatical connectivity’ another thought. In fact, no one does, but that’s another matter. Similarly, it doesn’t occur to the authors that absence of first-person singular (clauses with ‘I’) might be because the informants want to avoid confusion as to who is speaking. Or that they use indirect speech. Once again – the assumption that their informants are simply mouthpieces for the voices they hear makes very little sense!

The conclusion to this blogpost is bitter. Some time ago, I wrote a post which arose out of frustration with banging my head against the wall of medical journal (Running into a wall). I am tired of having papers rejected on the basis of spurious assumptions about language and communication. It’s very disheartening. But what’s even more disheartening is to see articles published with assumptions which fairly basic understanding of how communication works would dismiss. People are not mouthpieces, reports of something are not that thing, content is not linguistic. And I have read more of such papers than I care to remember, just as I have had more papers rejected by medical journals than I care to remember (note, I am not saying at all that my papers are good – I just want them to be assessed on the basis of linguistic knowledge and not on figments of medical assumptions about language).

Any further part of the conclusion would probably cross the lines of allowable academic frustration.