Literature week 6
Appraisal
On the scientific status of cognitive appraisal models of anxiety disorder – Richard J.
McNally
https://www-sciencedirect-com.proxy-ub.rug.nl/science/article/pii/S0005796700000735?via%3Dihub
Abstract
The cognitive paradigm for understanding and treating anxiety disorders comprises two distinct and
potentially incompatible approaches: appraisal and information-processing. Advocates of the latter
approach have sharply criticized the scientific adequacy of the appraisal models popularized by
cognitive therapists. The purpose of this essay is to provide a reappraisal of these critiques of
appraisal, and to defend an argument for methodological pluralism.
What passes for a coherent cognitive perspective actually comprises two sharply divergent and
potentially incompatible ways to understand aberrant cognition. One approach holds that emotional
disorder results from problematic beliefs and appraisals ascertainable via introspective self-report.
The other approach holds that emotional disorder results from information-processing abnormalities
whose influence is inferred only from overt behavioral indices (e.g., reaction times), not from self-
report. The tendency for many cognitivists to draw from both traditions masks a potentially fatal flaw
“that now seriously threatens the scientific credibility” of the cognitive paradigm. The purpose of this
essay is to evaluate the cogency of these concerns, especially in reference to cognitive approaches to
anxiety disorder.
1. The appraisal approach
When people refer to “the” cognitive approach to emotional disorders, they usually mean the
tradition inaugurated by Beck (1976). The central theme cutting across all these formulations is that
beliefs matter. Beliefs have consequences. Beliefs are not merely the epiphenomenal smoke of fires
burning elsewhere. Beliefs provide the causal basis for the mistaken appraisals of bodily sensations,
intrusive thoughts, and so forth that presumably underlie anxiety disorders. Counterfactually, if it
were not for certain problematic beliefs and their resultant catastrophic misappraisals, there would
be no disorder. Change the beliefs, and eliminate the disorder.
Accordingly, one formulates this type of cognitive model by ascertaining the relevant beliefs and the
resultant appraisals. Thus, Clark (1986) holds that panic attacks occur when a person misinterprets
certain bodily sensations as harbingers of imminent catastrophe, as exemplified by someone who
mistakenly believes that heart palpitations signify impending cardiac arrest.
Another prominent example is Salkovskis’s (1999) appraisal model of OCD (N.B.: he terms it a
“cognitive-behavioural” model, but appraisal is the central process). According to Salkovskis,
intrusive thoughts are converted into obsessions only when people interpret the thoughts in certain
ways. Beliefs about exaggerated responsibility are also deemed relevant.
,The appraisal approach is notably rationalistic. Anxiety disorders make sense as soon as one has
divined the relevant beliefs and resultant appraisals.
Further implied is the corollary of normalization of psychopathology. If the signs and symptoms of
disorder are expectable consequences of problematic beliefs, even neurobiological “abnormalities”
are viewable in a different light. For example, patterns of cerebral metabolic rates for glucose in
patients with OCD differ from those in healthy control subjects, and they tend to change following
either successful behavior therapy or pharmacotherapy.
2. Critique of the appraisal approach
Despite its track record for inspiring effective treatments, the appraisal approach has provoked sharp
criticism from other cognitivists. Thus, Lang (1988) argues that it rests on “an inherently limited level
of discourse” (p. 221), noting that: It has the force of a good story, and does not ask us to believe in
any cognitive mechanism beyond those that have been familiar to playwrights and novelists for
centuries. Above all, it can be readily comprehended by the patient, whose cooperation can then be
enlisted in lending substance to the scenario.
According to Lang, to say that someone consciously (or unconsciously) appraises bodily sensations as
harbingers of imminent catastrophe, and then to attribute the subsequent panic attack to this
misappraisal is to explain panic in the same way that Molie´re’s (1950) “imaginary invalid” explained
the sleep-inducing effects of opium by invoking its dormitive virtue. Clark’s (1986) cognitive model of
panic “has, to say the least, similar explanatory limitations”.
MacLeod (1993) is likewise critical of the appraisal approach. Because of its reliance on introspective
self-report to ascertain beliefs, the appraisal approach embodied in cognitive therapy falls outside
“the boundaries of legitimate science”.
According to MacLeod, the cognitive revolution in experimental psychology involved a liberalization
in theory, but not in method.
In striking contrast, MacLeod observes, the cognitive revolution in clinical psychology, initiated by
cognitive therapists, embodied a liberalization in method as well as theory. It licensed requests of
patients to consult their phenomenology and furnish self-reports about their “cognitions,” thereby
forfeiting claim to the mantle of rigorous science.
Citing Nisbett and Wilson (1977), MacLeod stressed that people are often very poor at identifying the
factors influencing their behavior.
The search for causes is understandable. Baffling emotional events, such as spontaneous panics,
prompt a search for explanations, and this “effort after meaning” may beguile patients (and
clinicians) into thinking they have hit upon the relevant causal variables. For example, merely
because catastrophic thoughts occur between detection of bodily sensations and panic does not
confirm causality.
In addition to drawing on Nisbett and Wilson’s work, MacLeod (1993) furnished examples from
cognitive psychology showing that people are often unaware of the causes of their behavior, thereby
undermining the value of their introspective explanations.
Not only is introspective self-report inadequate for revealing causes that lie outside of awareness, it
fails to capture processes that occur very quickly.
,Finally, with respect to the normalization corollary implicit in the appraisal approach, it is important
to be clear about what is being normalized in anxiety disorders and in psychopathology more
generally. According to Maher, delusions are rational attempts to make sense of anomalous
experiences such as Schneiderian first-rank symptoms of thought broadcasting, experiences of
influence, and auditory hallucinations. Firstrank symptoms are phenomenologic events caused by
deviant processes in the brain. One can “normalize” delusions by noting that they reflect rational
appraisals of inexplicable pathological events. Likewise, while attempting to normalize catastrophic
appraisals in the anxiety disorders, we must be alert to the possibility that the events so appraised
may themselves arise from deviant brain processes. Appraisals may be rationally explicable, but that
does not necessarily mean that the events (e.g., intrusive thoughts, bodily sensations) are likewise
“normal.”
3. Reappraising critique of appraisal
Critiques issuing from either inside (e.g., MacLeod, 1993) or outside (e.g., Klein & Klein, 1989) the
cognitive paradigm have not slowed the appraisal approach, especially in terms of treatment
development. The appraisal approach has been markedly more therapeutically fruitful than has the
information-processing approach.
But merely because a theory inspires effective treatments does not confirm its truth.
In addition to its therapeutic fruitfulness, the appraisal approach has another source of staying
power: folk psychology. Folk psychology is the quotidian theory we apply in everyday life to predict
and explain the behavior of others. Without shared intuitions about belief, desire, and so forth it
would be impossible to make head or tail out of the practices of diverse cultures around the globe.
However, most critiques of the appraisal approach do not concern either therapeutic issues or
eliminative materialist objections to folk psychology as such. Most critiques address presumptive
conceptual and methodological flaws. One conceptual concern, expressed by Lang (1988), is that
explanation in terms of beliefs has all the limitations of the “imaginary invalid’s” account of the
sleep-inducing properties of opium. Presumably, invoking a catastrophic interpretation as the cause
of panic begs the question about panic’s causes just like adducing the dormitive virtue begs the
question about why opium puts people to sleep.
Invoking dormitive virtues (i.e., a conceptual placeholder for a genuine sleep-producing causal power
present in opium, but lacking in drugs like amphetamine) and catastrophic misinterpretations for the
causes of sleep and panic, respectively, is the beginning, not the end, of inquiry. In fact, David M.
Clark himself recently expressed the hope that ecologically valid cognitive-experimental methods will
be used to “delineate the subtle processing biases that maintain sufferers’ fears”.
MacLeod is unquestionably correct that overt behavioral indices are the sine qua non of
measurement in cognitive psychology. But self-report is not wholly disreputable, not even in
cognitive psychology. Often self-report data provides the foundation for a research program that
itself does not rely on self-report (e.g., classifying high- and low-anxiety groups by letting them make
self-reports questionnaires).
MacLeod provided several examples of research whereby subjects’ behavior was affected by
influences which they were unable to report. Ironically, the point he (correctly) makes rests on
imputation of consciousness or the lack thereof. Other areas of cognitive science also rely on
phenomenal awareness. Introspective self-reports provide the database.
, Tulving (1985) has been especially forthright about how phenomenal awareness must figure in
cognitive science. He argued that progress in understanding human memory requires “the
postulation and identification of different kinds of consciousness but also their measurement as an
aspect of experience, or as a dependent variable” (p. 2, emphasis in original). Tulving argued that
people in recognition memory experiments use two bases for making recognition judgments: know
vs remember. For example, recognizing that one has seen a particular word earlier in the experiment
can be done on the basis of phenomenal recollection (e.g., remembering other thoughts or feelings
triggered by the word) or simply knowing — being confident that one must have seen it — without
having a recollective experience. Tulving argued that remember judgments were based on
autonoetic awareness, a form of consciousness integral to autobiographical (episodic) memory,
whereas know judgments required only noetic awareness, a form of consciousness integral to
semantic memory. Tulving’s group has endeavored to anchor these forms of consciousness to the
functional anatomy of the brain.
In some cases, self-reports may actually be isomorphic with the cognitive mechanisms themselves, as
in the think-aloud problem-solving protocols of Herbert A. Simon. Finally, many aspects of
psychopathology do not have overt behavioral manifestations other than self-reports of phenomenal
states.
4. Conclusions
Taken together, the arguments and counterarguments about self-report, introspection, and appraisal
reviewed here strongly indicate that a methodological pluralism is warranted. Sometimes self-reports
are just as useless and misleading as the critics charge. But sometimes, depending on one’s question,
they can be precisely what is needed. If one seeks to divine a patient’s beliefs, the patient’s self-
report is essential. Likewise, certain phenomena, such as obsessions, have no outward
manifestations other than that revealed in language. Even PET studies of OCD patients require one to
confirm that the person is, indeed, obsessing in the scanner. On the other hand, these methods are
useless at best and misleading at worst for other kinds of questions. People can often provide
answers to what-questions, but often cannot provide answers to how-questions. In summary, rather
than expel self-report from the study of anxiety disorders by fiat, we should attend to when it
provides the answers we seek and when it does not.
