Discourse planning
One possibility is that psychotic people have difficulty planning what they are going to say. Consistent with this idea, Ralf Hoffman, at Yale University in the USA, inferred ineffective discourse planning in thought-disordered patients from a structural analysis of their speech, in which the links between different ideas were traced. It was found that propositions contained in the speech of ordinary people seemed to be organized into a ‘strong hierarchy’ of ideas whereas this was not the case for the speech of thought-disordered schizophrenia patients.35 In a subsequent study, Hoffman claimed that his method allowed him to discriminate between the disordered speech of schizophrenia
Table 15.3 Categories of linguistic abnormality measured by the Communication Disturbance Index (Docherty, DeRosa and Andreasen, 1996).
Category
Definition
Example
Vague references
Over-inclusive words or phrases that obscure meaning because of a lack of specificity.
‘Being is is, it’s not bad. You can do things and plus you can make people afraid of you.’
Confused references
Words or phrases that refer ambiguously to one of at least two clear-cut alternative referents.
‘Take the clock, for instance. You got ten, twelve on it, you got other numbers on it, you got a volume button on it, it go up and down.’
Missing information references
References to information not previously presented and not known to the listener.
‘I like to work all right. Some of those shops were filthy… I liked the bakeries, some of the shops are clean.’ (No prior mention of shops or bakeries.)
Ambiguous word meanings
Instances in which a word or phrase could have a number of different meanings, and the intended meaning is not obvious from the context.
‘I had a chance to grow with him but I got a divorce because I couldn’t.’
Wrong word references
Use of an odd or apparently inappropriate word or expression in an otherwise clear utterance.
‘I was trying to predict them people that I need, I need to get out of there.’
Structural unclarities
Failures of meaning due to a breakdown or inadequacy of grammatical structure.
‘I was socializing with friends. Girlfriends and friends the same as male.’
patients, and that of manic patients, which was characterized by rapid shifts from one discourse plan to another.36 Unfortunately, Hoffman’s method of analysis has been criticized by other researchers for being unreliable and subjective.37
Although other researchers have reported findings that have been interpreted (to my mind, unconvincingly)as consistent with Hoffman’s theory,38 it seems fatally flawed for reasons that we identified in the last chapter. The idea that speech requires a plan in order to be coherent raises the spectre of an infinite regress in which a plan is required in order to compose the discourse plan, and seems to underestimate the extent to which people can construct coherent speech ‘on line’.
Speaking to the needs of the listener
A second possibility is that thought-disordered patients are somehow unable to adjust their speech to meet the needs of the listener. In principle, this might happen for one or both of two reasons. First, patients may be unable to monitor their own speech, so they fail to recognize that it is unintelligible. Second, they may be unable to understand what the needs of the listener are.
The study by Martin Harrow and Joan Miller which I discussed earlier is consistent with the first of these hypotheses; it will be recalled that thought-disordered patients failed to recognize that their own speech was unintelligible to others but could easily recognize when the speech of others was incoherent. An observation that is also consistent with the hypothesis was made by linguist Elaine Chaika when she asked psychotic patients and ordinary people to repeat a short story; when digressing or breaking off in the middle of a grammatical construction the normal participants would signal that they had done so and then return to the story, whereas the schizophrenia patients appeared unaware they were deviating from the task.39
A more sophisticated approach to examining the role of self-monitoring in thought-disordered patients was used in a series of studies conducted by Phil Harvey at Mount Sinai Hospital in New York. In the last chapter I introduced the concept of source monitoring when discussing hallucinations, focusing on difficulties that patients experience when attempting to distinguish between things that they hear and things that they think. However, there is another type of source monitoring in which we distinguish between what we have thought and what we have said. Clearly, an impairment of this skill will make individuals vulnerable to omitting important segments of information when speaking to others, making their speech appear incoherent.
Throughout the 1980s, Harvey and his colleagues carried out a number of studies in which patients were asked to read one set of words aloud, and to imagine themselves reading a second set of words (the ‘think’ condition). Afterwards, they were shown a list of words including those they had read aloud, those they had thought, and some which had not been presented previously, and were asked to say which of them they had read and which of them they had thought. In general, it was found that poor performance on this kind of task was associated with thought disorder in schizophrenia patients but not in manic patients.40
The second possible explanation for the thought-disordered patient’s inability to speak to the listener’s needs has also received some support. If thought-disordered patients cannot understand those needs, we should expect them to perform badly on theory-of-mind tests. This possibility has been explored in a series of studies recently undertaken at the Centre Hospitalier de Versailles in France by Yves Sarfati, Marie Christine Hardy-Bayle and others.41 In brief, these studies have shown that patients who have high scores on Andreasen’s TLC Scale perform less well on a variety of ToM tasks than patients who show no evidence of thought, language and communication disorder.
Working memory
A third possibility is that more basic cognitive processes are compromised in thought-disordered patients. We have already seen evidence that these patients suffer from deficits in working memory (the part of the memory system responsible for holding information for brief periods of time). In Chapter 8, I described a study by Tom Oltmanns and John Neale in which it was discovered that poor performance on the digit span with distraction task was associated with communication disorder.42 The findings of Grove and Andreasen, which we considered earlier, are also consistent with this observation, as are the results of many other studies.43
In attempts to demonstrate a causal connection between working memory deficits and communication disorder, some investigators have experimentally impaired the working memory system in ordinary people. Deanna Barch and Howard Berenbaum of the University of Illinois at Urbana-Champaign interviewed undergraduate students in two conditions: under normal circumstances and in a ‘dual-task’ condition in which they also had to perform a task that was designed to keep the working memory system busy. As a consequence, the speech of the students showed less grammatical complexity and more long pauses. As Barch and Berenbaum note, these characteristics have been observed in the speech of psychotic patients. However, other well-documented features of psychotic speech (for example, incompetent references) were not observed.44
More dramatic impairments of working memory can be temporarily induced by administering ketamine, and a number of investigators have therefore studied the speech of healthy volunteers who have been persuaded to take this drug. A group of investigators at the US National Institute of Mental Health gave an infusion of ketamine to ten ordinary people, who, afterwards, showed evidence of thought disorder as measured by Andreasen’s TLC Scale, and also poor performance on a working memory measure. Interestingly, the decrease in working memory performance experienced by the participants correlated with the severity of their communication dif
ficulties.45
Terry Goldberg and Daniel Weinberger, also at NIMH, have pointed out that the relationship between working memory and communication disorder may not be as simple as these results seem to suggest. Neurological patients who have damage to their frontal lobes often suffer from impaired working memory, but almost never show evidence of thought disorder. On the other hand, thought-disordered patients often respond abnormally on simple word association tests, which do not require the use of working memory.46
Was Bleuler right?
A final approach to understanding the relationship between cognitive deficits and communication disorder has focused on a part of our cognitive system known as semantic memory, and bears some similarity to Bleuler’s original theory. The semantic memory system – which stores meanings, knowledge and ideas – is thought to consist of complex networks of concepts, connected by associative links. According to this very general model (of which there are many specific versions), one concept leads to another by a process of ‘spreading activation’ – as a concept becomes awakened in the mind (because the individual is thinking about it) closely connected concepts are also activated or ‘switched on’. In this way, one idea leads to another.
A phenomenon that demonstrates this effect is known as ‘semantic priming’. In a semantic priming experiment, participants are shown words in quick succession and are asked to indicate in some way (for example, by reading aloud, or by pressing a button on a computer) when they have recognized each word. Recognition is faster (by a few tens of milliseconds) if a word is preceded by a meaningfully related word (for example, if ‘black’ is preceded by ‘white’) than if it is preceded by an unrelated word (for example, ‘soft’).47 The conventional explanation for this effect is that perception of a word (‘black’) leads automatically to the partial activation (or ‘priming’) of related words (‘white’), so that the brain is already ‘tuned’ into them, and recognition is easier.
It is believed that many of the concepts represented in semantic memory are organized hierarchically. For example, the concept of ‘living things’ includes ‘animals’, which includes ‘birds’, which in turn includes ‘crows’. A robust finding from ordinary people is the so-called category-relatedness effect– it takes us less time to decide that a typical exemplar belongs to a category (‘a crow is a bird’) than it takes us to decide that an atypical exemplar belongs to a category (‘a penguin is a bird’). Again this can be understood in terms of activation spreading from the category concept to typical exemplar concepts, but not to the atypical exemplars.
As we have already seen, some of the earliest attempts to study these processes in schizophrenia patients were made by Bleuler and Jung, who examined patients’ word associations. In fact, modern research has consistently shown defective semantic memory in schizophrenia patients. For example, some studies have shown that patients often perform poorly when asked to sort objects into categories, to remember object names, or to describe the defining features of concepts such as ‘animals’ and ‘birds’.48 Although there is evidence that these problems cannot simply be attributed to poor motivation or more general intellectual deficits, many of these studies have been compromised by the failure to link the findings to particular complaints. Over the past few years, however, evidence has accumulated supporting Bleuler’s view that abnormal semantic processes play a specific role in disordered communication.
Some of this evidence has emerged from studies of word associations. Manfred Spitzer at the University of Heidelberg in Germany replicated some of Bleuler’s original experiments, and confirmed that thought-disordered schizophrenia patients show fewer typical associations and more indirect or ‘mediate’ associations.49 In a study by Terry Goldberg and his colleagues in the USA, a group of schizophrenia patients and a group of normal participants were given a battery of psychological tests. Two measures discriminated those patients with severe thought disorder from those with mild or no thought disorder – a verbal fluency test (in which the individual is asked to think of as many words as possible beginning with a particular letter) and a vocabulary test – both of which depend heavily on semantic memory.50 Goldberg and his colleagues went on to use patients’ responses on a test in which they had to generate words belonging to particular categories, to ‘map’ the way that their semantic systems were organized. For example, by analysing the relationships between successive words generated during the task, it was possible to show that most ordinary people organize the category ‘animals’ along two dimensions: wild–domestic and large–small. However, when maps were generated for thought-disordered patients, they appeared to be less well organized than those of non-thought-disordered patients or ordinary people.51
Less consistent findings have been obtained when researchers have attempted to assess semantic functioning by measuring the kinds of priming effects I described earlier. Brendan Maher in the United States52 and Manfred Spitzer in Germany53 have reported that priming is increased in patients with communication disorder. In a particularly complex experiment, Spitzer has also observed abnormal spreading activation effects.54 Whereas priming of direct associations (for example, priming of ‘foot’ by ‘shoe’) was found in ordinary people, communication-disordered patients also showed priming for indirect associations (for example, ‘hand’ by ‘shoe’, presumably mediated by ‘foot’).
Other studies, for example by Terry Goldberg and his colleagues,55 have observed reduced priming in patients with thought disorder. It is possible that some of these inconsistencies are more apparent than real. Priming effects are notoriously sensitive to experimental conditions – the duration for which words are presented, and the interval between them, for example. One possible interpretation of the evidence is that reduced spreading activation to immediate associates, and increased activation of indirect associates – either as separate processes or the consequence of a single abnormality in the semantic memory system – make it more difficult for the thought-disordered patient to find appropriate words when speaking.
The Roots of Incoherence
It is possible that dysfunctions in working memory, source monitoring, theory of mind and semantic memory are all implicated in thought disorder to some degree. As American psychologists Deanna Barch and Howard Berenbaum have noted, language production processes are highly complex, and it is therefore likely that abnormalities in a number different cognitive mechanisms can influence them.56
A tentative model that tries to draw together some of the ideas we have considered in this chapter is shown in Figure 15.2. It suggests that semantic problems, which I assume to be long-standing, make individuals vulnerable to thought, language and communication disorder. It further assumes that, in vulnerable individuals, emotional arousal leads to working memory deficits, which in turn lead to poor source monitoring and ultimately to speech that is incomprehensible to others. However, I must emphasize that this account is highly speculative (and certainly not secure enough to be deemed a ‘theory’), so it is possible that I have misjudged the flow of causal relationships between the different processes that appear to influence psychotic speech. More research is surely needed.*
Figure 15.2 A very tentative model of positive thought, communication and language disorder.
One remaining matter of controversy is the relationship between communication disorder in schizophrenia patients and communication disorder attributed to mania. Some researchers have suggested that these are quite different, whereas others have assumed that they are similar or even identical. My own prejudice – fuelled by my earlier critique of the Kraepelinian paradigm – is to consider them similar. However, on more careful consideration it becomes obvious that ‘similar’ or ‘different’ do not exhaust the possibilities. In a thoughtful discussion of these issues, Linda Grossman and Martin Harrow remarked:
We believe that the disordered cognition that is so prominent in manic and schizophrenic patients is the product of several underlying factors acting together, rather than a single factor
. Further, it is probable that there are underlying factors leading to psychopathology that are common to both manic and schizophrenic patients, and that there are also other underlying factors that differ in these two patient groups… Until recently, the assumption that thought disorder is found only in schizophrenia has so governed diagnostic practice as to shape routine clinical observation to such an extent that many grossly disorganized manic patients who thought and verbalized strange ideas were not seen as having disordered cognition.57
Part Four
Causes and their Effects
16
Things are Much More Complex than they Seem*
My own suspicion is that the universe is not only queerer than we suppose, but queerer than we can suppose.
J. B. S. Haldane1
In the previous parts of this book, I have attempted to show how psychological research has begun to unravel the processes involved in the various complaints of psychotic patients. In each case, these experiences and behaviours appear to be understandable when viewed in the context of what we know about the normal human mind. It is important to repeat that there is nothing ‘anti-biological’ in this analysis. The psychological processes that give rise to complaints are located in the brain (although whether psychosis is caused by damage to the brain is another matter, which we will return to later).
Madness Explained Page 46