Tuesday, 12 January 2016

Response to McKay's 'Bayesian Accounts and Black Swans'

This post is by Matthew Parrott (pictured below) and Philipp Koralus (pictured below). In a previous post they summarised their recent paper ‘The Erotetic Theory of Delusional Thinking’, published in Cognitive Neuropsychiatry. Ryan McKay, in a post the following week, summarised his recent paper responding to Matthew and Philipp, 'Bayesian Accounts and Black Swans: Questioning the Erotetic Theory of Delusional Thinking'. In this post, Matthew and Philipp respond to Ryan. 



We are very grateful to Ryan McKay for taking the time to read our paper, and for formulating helpful questions in response to it. We hope that by addressing these questions, we can clarify any aspects of our theory that might be puzzling.


1) How can we reliably distinguish exogenous and endogenous question raising?

As we said in our original post, we propose that patients ‘entertain roughly the same default questions that most people strongly associate with various external stimuli, but that they either envisage fewer alternative possible answers to these questions or raise fewer follow up questions’. Thus, according to our theory, there is a difference between exogenously raising a question, which is when the process of raising a question is the result of some external stimulus, and endogenously raising a question, which is when a patient raises a question independently of external stimuli. This sort of distinction is familiar in cognitive science. We find it, for example, in discussions of attention (i.e., endogenous attention is internally controlled and exogenous attention is externally controlled). One would presumably differentiate these experimentally in roughly the same way that one would for attention, by controlling for external stimulus.

2) If deluded individuals are selectively deficient in raising their own questions, why are they unable to fully utilize and retain questions that others raise?

In our paper, we propose that one virtue of the Erotetic Theory is that it can actually explain why many delusional individuals come to momentarily doubt their delusion in response to external questioning but then quickly fall back into it. As we note in our paper, an individual would have to take the right questions on board while evaluating their perceptual experience. Just posing more questions externally after that wrong conclusion has already been drawn may not be enough (and does not seem to be enough in practice). The study by Breen and collaborators that McKay mentions is completely consistent with this picture. One might also note that MF appears to acknowledge that his delusion is ‘strange’ during the course of a clinical interview, which we think shows that he is not ‘completely impervious to external questioning’ as McKay suggests. In any case, it is not terribly surprising that MF’s delusion is relatively insensitive to counterevidence or to general considerations of plausibility, which both seem to be general features of delusional cognition. But it is worth emphasizing that raising a question in the technical sense of the Erotetic Theory is not the same as presenting a person with some evidential considerations against something he or she believes, after a delusional conclusion has already been drawn.

3) If I encounter a person lacking the C-feature, what could lead me to speculate that the person is an imposter?

McKay seems to think that representing one’s spouse as lacking what we call a C-feature is not sufficient to speculate that the person is an imposter. We disagree. As we claim in the paper, one’s spouse is normally represented as a set of features, one of which is something like ‘intimacy’ (represented as a property of the person), or what we discuss as the ‘C-feature’ in our paper. If one’s representation lacks this, then one will answer the question: ‘who is this among the people I know?’ with ‘no one that I know’, i.e., a stranger. Notice that this means that one will have a conscious visual experience of someone who looks exactly like one’s spouse but one will also think the person is a stranger (no one that I know). That is almost a definition of an imposter.

4) If the absence of the C-feature prompts imposter speculations when I encounter someone I know, why would the absence of other relevant features (e.g., a moustache that has been shaved off) not do the same?

We suspect that people process some properties as essential and some as superficial. We suspect that there is only a serious possibility of robust misidentification if we mistakenly think that an essential property is lacking. We proposed that the C-feature, as we called it, is taken as an essential property of a close loved one, unlike, say, having a moustache. A key difference may be that you can reconcile lack of a superficial property with identity without having to envisage the possibility of misperception on your own part.

5) Why would the suppositional reasoning strategy require a capacity for endogenous-question-raising, whereas the correct logical strategy wouldn’t?

According to the Erotetic Theory, suppositional reasoning is formally defined as an inquisitive process. Nothing about being presented with a conditional reasoning task requires someone to adopt a suppositional strategy; this would require one to endogenously raise a question. As the paper demonstrates, an extremely interesting consequence of this is that suppositional reasoning generates errors in certain reasoning tasks.

6) Is the Erotetic Theory a theory of delusions, or schizophrenia, or both?

As we tried to explain in our original post, the Erotetic Theory is a theory of reasoning. In our paper, we demonstrate how this theory can both explain and predict different patterns of reasoning that we see exhibited by delusional subjects, namely the tendency to ‘jump to conclusions’ on probabilistic reasoning tasks and the adoption of the Capgras delusion in response to highly anomalous experiences. The theory also explains a pattern of reasoning exhibited by schizophrenic subjects (Mellet, et al; 2006).

At this stage, it is probably best not to think of our theory as a general account of delusion or schizophrenia, but rather as a proposal about one cognitive deficit that can lead to the patterns of reasoning observed in psychiatric subjects. It may be that this deficit is strongly associated with delusion (schizophrenic subjects are highly prone to delusion), or with schizophrenia (schizophrenics exhibit the JTC bias and it is also an aetiology of Capgras delusion). We think this is an interesting question that is worth pursuing.

7) What is the equivalence between raising fewer questions and raising lesser questions (questions with fewer alternatives)?

On our regimented account, questions are sets of alternatives. Aggregating multiple sets of alternatives (i.e., multiple questions) into one larger set of alternatives, and just taking on board a question with more alternatives largely come to the same thing.

8) Does the suggestion that individuals are less inquisitive add to what we know (which is that certain subjects seek less evidence when forming beliefs and making decisions)?

The Erotetic theory gives us an explanation of why subjects seek less evidence. It also makes sense of phenomena that are not obviously captured by bare notions of “seeking less evidence,” for example the results on reasoning with conditionals.

No comments:

Post a Comment

Comments are moderated.