ROBERT MICHELS
519
apist and was implanted there by the theory that guided the inquiry. I would
say this is true. It is a major discovery of psychoanalysis. Crews doesn't go
into this, but it is, of course, true of all therapies. It isn't that psychoanalyt–
ic therapy is vulnerable to suggestion and there are therapies that aren't, it's
that psychoanalytic therapy is almost unique in its immense concern about
the possibility of suggestion, its attempt to account for suggestion in the
process of treatment, and in its curiosity about the basis for suggestion. In
a sense the difference between psychoanalysis and many other psychother–
apies is that other psychotherapies work by suggesting healthy behavior,
while psychoanalysis works by exploring the patient's attempt to transform
the therapy into suggestion. Instead of exploiting the patient's willingness
to receive suggestions, psychoanalysis uses that willingness to liberate the
patient from his or her dependency on external suggestions. Psychoanalysis
is different from the other therapies in that it tries to use suggestibility as a
lever for freeing people from their vulnerability to suggestion rather than
exploiting it as a lever for helping people. Does that make it immune to the
danger of suggestion? No. There is no way to make it immune. Is it possi–
ble that a well-meaning analyst will inadvertently make suggestions that
patient and analyst will later believe to be true? Of course it is possible. Is
there any way to guarantee that that doesn't happen and that the data are
"safe" and guarded from all possible contamination? No way that I know
of. What we are saying is that the data of psychoanalysis, like that of all other
therapies we know, are susceptible to the contamination of suggestion. It is,
unlike other therapies, preoccupied with trying to reduce that contamina–
tion. But there is no assurance that it can be totally successful at that.
What does this mean for the status of psychoanalysis? For Griinbaum
the problem of suggestibility is virtually a fatal flaw destroying its poten–
tial as a natural science. This involves several assumptions. One is that the
data we are talking about are the data of statements made in analytic ses–
sions. Is it essential that those statements are scientific hypotheses that can
be proved or disproved like hypotheses in physics or chemistry if psycho–
analysis is to be a science? Grunbaum seems to believe that it is and Crews
accepts Griinbaum's position that if we can't test the scientific validity of
interpretations made in analytic sessions then analysis is not scientific.
There are other points of view. One would challenge that particular
definition of science. Griinbaum is borrowing from Karl Popper who
believed that science required the falsifiability of hypotheses. This is an
important view, but it is not the only view of what science is. It is not even
the most popular view today. An alternative would define science as some–
thing that generates an interesting strategy for discovery, inquiry, learning,
and validation through models or theories or hypotheses. The notion of
falsifiability as the core criteria is too narrow a concept of science. Still
