Available online at www.sciencedirect.

com

Studies in History
and Philosophy of
Biological and
Biomedical Sciences
Stud. Hist. Phil. Biol. & Biomed. Sci. 39 (2008) 5264
www.elsevier.com/locate/shpsc

On the validity of Freuds dream interpretations

Michael Michael
Underwood International College, Yonsei University, 134 Shinchon-dong, Seodaemun-gu, Seoul 120-749, South Korea
Received 18 September 2006; received in revised form 27 March 2007

Abstract
In this article I defend Freuds method of dream interpretation against those who criticise it as involving a fallacynamely, the
reverse causal fallacyand those who criticise it as permitting many interpretations, indeed any that the interpreter wants to put on
the dream. The rst criticism misconstrues the logic of the interpretative process: it does not involve an unjustied reversal of causal
relations, but rather a legitimate attempt at an inference to the best explanation. The judgement of whether or not a particular interpretation is the best explanation depends on the details of the case in question. I outline the kinds of probabilities involved in making the
judgement. My account also helps to cash out the metaphors of the jigsaw and crossword puzzles that Freudians have used in response to
the many interpretations objection. However, in defending Freuds method of dream interpretation, I do not thereby defend his theory
of dreams, which cannot be justied by his interpretations alone.
2007 Elsevier Ltd. All rights reserved.
Keywords: Sigmund Freud; Dream interpretations; Reverse causal fallacy; Inference to the best explanation

When citing this paper, please use the full journal title Studies in History and Philosophy of Biological and Biomedical Sciences

1. Introduction
Freuds critics have made numerous objections to his
method of interpreting dreams. One of these is that the
method permits, not just one interpretation, but innitely
many of them, indeed just about any interpretation that
the interpreter wants to put on the dream. If, for example,
one wants to interpret a dream as wish-fullling, it seems
easy enough to nd some justication for associating a
wish with the content of the dream. A second objection is
that the method involves a fallacy: the reverse causal or free
association fallacy. While it may be true that the material
that emerges during free association is determined by the

E-mail address: michael_t_michael@yahoo.co.uk

1369-8486/$ - see front matter 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.shpsc.2007.12.014

content of the dream, it is fallacious to infer that this material therefore determined the content of the dream.
Clark Glymour raises both of these objections in his
article The theory of your dreams. He raises the rst by
arguing as follows:
we know before-hand that, if a person is asked to associate his thoughts with elements of a dream, and report
his associations, after a while we will be able to make up
a cogent story, thought, fear, wish, or whatever from
the resulting associations. . . There is nothing special
about dreams in this regard; much the same could be
done with rock formations or with blotches of ink or

M. Michael / Stud. Hist. Phil. Biol. & Biomed. Sci. 39 (2008) 5264

Thematic Apperception Tests or whatever. Unless he or

she is enormously stupid, the person doing the associating can make up such a story from elements of his or her
associations. Indeed if the associater is clever, many
such stories can be made up. (Glymour, 1983, p. 60)
Freud would of course have denied that he was making
up stories from the material uncovered during the dreamers free associations. On the contrary, he claimed that
the interpretations were logically compelled by the material. According to him, interpreting a dream is like solving
a jigsaw puzzle, where once one is able to put some of the
pieces together, an inkling of the overall picture emerges
that helps one to piece together the rest of the puzzle. When
one has completed the puzzle the picture acquires a meaning, so that there is no gap anywhere in the design and so
that the whole ts into the frameif all these conditions
are fullled, then one knows that one has solved the puzzle
and that there is no alternative solution (Freud, 1923/
1961, p. 116). Alternatively, interpreting a dream is like
solving a crossword puzzle, where once some words are
in place, with skill someone can solve the whole puzzle.
When one nds a solution of a particular clue that also
happens to t in with the letters already there, one can infer
that this is the correct solution, and in turn, the letters contributed by this solution help to nd a solution to the next
clue. These metaphors are useful, but we need to describe
what they are metaphors of if we are to make a proper
rebuttal of Glymours objection.
Next Glymour raises the objection that Freuds method
depends on the free association fallacy. He does this in the
context of an example of one of Freuds interpretations.
The dream in question, by a woman patient, is as follows:
I wanted to give a supper-party, but I had nothing in the
house but a little smoked salmon. I thought I would go
out and buy something, but remembered then that it was
Sunday afternoon and all the shops would be shut. Next
I tried to ring up some caterers, but the telephone was
out of order. So I had to abandon my wish to give a supper-party. (Freud, 1900/1991, p. 299)
The woman was then asked to produce associations based
on this dream, and eventually she came up with the
following:
After a short pause, such as would correspond to the
overcoming of a resistance, she went on to tell me that
the day before she had visited a woman friend of whom
she confessed she felt jealous because her (my patients)
husband was constantly singing her praises. Fortunately
this friend of hers is very skinny and thin and her husband admires a plumper gure. I asked her what she
had talked about to her thin friend. Naturally she
replied, of that ladys wish to grow a little stouter. Her
friend had enquired too: When are you going to ask
us to another meal? You always feed one so well.
The meaning of the dream was now clear, and I was able
to say to my patient: It is just as though when she made

53

this suggestion you said to yourself: A likely thing! Im

to ask you to come and eat in my house so that you may
get stout and attract my husband still more! Id rather
never give another supper-party. What the dream was
saying to you was that you were unable to give any supper-parties, and it was thus fullling your wish not to
help your friend to grow plumper. The fact that what
people eat at parties makes them stout had been brought
home to you by your husbands decision not to accept
any more invitations to supper in the interests of his plan
to reduce his weight. All that was now lacking was some
coincidence to conrm the solution. The smoked salmon
in the dream had not yet been accounted for. How, I
asked, did you arrive at the salmon that came into your
dream? Oh, she replied, smoked salmon is my friends
favourite dish. I happen to be acquainted with the lady
in question myself, and I can conrm the fact that she
grudges herself salmon no less than my patient grudges
herself caviar. (Freud, 1900/1991, pp. 230231)
Glymour claims that the association the woman makes
with the dream, the thought of her skinny friend, is caused
by two elements of the dream, the supper-party and the
smoked salmonthat is, the causal links are as shown in
Fig. 1 and that Freud transposes the causal relation, to
give Fig. 2. This is a fallacy, for what are the grounds of
the transposition? Evidence for the rst causal model is
not necessarily evidence for the second (Glymour, 1983,
p. 68). Freuds most notable recent critic, Adolf Grunbaum, endorses Glymours argument wholeheartedly:
Indeed, Freud thus argues fallaciously from the conuence
of associations to a causal reversal in explicitly generalized
form. (Grunbaum, 1984, p. 234)
I believe that both the many interpretations objection
and the objection that Freuds method is based on the free
association fallacy are wrong. Freuds method of interpretation does not involve a fallacy and, when properly
applied, leads to (inductively) valid interpretations. However, in saying this I caution against an error that is easily
made: we must not confuse the epistemological status of
Freuds method of interpretation, as exemplied in the
numerous case studies he provides, and the epistemological
status of his theory of dreams. A basis for the above con-

supper-party

smoked salmon

Thought of skinny
friend
Fig. 1.

54

M. Michael / Stud. Hist. Phil. Biol. & Biomed. Sci. 39 (2008) 5264

Thought of skinny
friend

supper-party

smoked salmon
Fig. 2.

fusion would be the assumption that the interpretations are

intended to provide the only justicatory support for the
theory. This assumption is false. Freud himself did not
believe that he could prove the general theory that the instigator of dreams is always an infantile wish from his interpretations alone. He was aware, and did not hide the fact,
that the theory of dreams depended on some complex and
intricate theorising about the mind.1 One must therefore
approach the interpretations as demonstrating the method
of interpretation rather than having probative value for the
theory of dreams, though they are doubtless intended to be
theoretically suggestive and to some extent conrmatory of
the theory.2
2. Critique of Glymour
If we return to Glymours illustration of the free association fallacy we will see that the objection is not as devastating as it might rst appear. For one, Glymours
suggestion that there were two elements that caused the
thought of the skinny friend, both the supper-party and
the smoked salmon, is misleading, especially in the light
of his later assertion that one associates, at Freuds direction, until one thinks of something which has connections
with several elements in ones dream. There is no indication in Freuds account that the smoked salmon element
of the dream was a cause of the thought of the skinny
friend.3 It is therefore misleading for Glymour to make this
suggestion, for by presenting the smoked salmon element
as a cause he diminishes without argument the conrma-

tory value of the subsequent discovery of a connection

between the skinny friend and smoked salmon. What
Freuds account actually amounts to is more like the following. Starting o from thinking about the dream, the
woman associated the thought of her skinny friend, so
we can say the dream was a causal inuence in her thinking
about the skinny friend. Freud then inferred that the
thought of the skinny friend caused the supper-party element of the dream, and by extension that thoughts about
the skinny friend were a cause of the content of the dream
in general. It was only after this theory had been formed
that Freud prompted the woman to explain the smoked salmon element, to which the woman immediately responded
by connecting the element with her skinny friend. That is,
she oered evidence that thoughts about the skinny friend
are likely to cause thoughts of smoked salmon. This then
conrmed the theory. That is, it supported the theory in a
way similar to that in which a successful prediction supports a theory, rather than to the way in which accommodated evidence supports a theory. As Peter Lipton argues
in Prediction and prejudice (1995), we have more reason
to believe a theory that successfully predicts data than
one that accommodates it. Similarly, if the above analysis
is correct then we have more reason to believe Freuds theory than if Glymours analysis is correct.
There is however a more serious aw in Glymours argument. This is his failure to adequately follow up a suggestion that he himself makes, that Freuds method in this
case involves, not an arbitrary transposition of a causal
relation, but rather an inference to the best explanation.
He makes this suggestion with the following comments:
If we think simply in terms of the second causal picture,
the fact that the dream both contains a failed attempt at
a supper-party and mentions smoked salmon, and both
of these elements lead to the remembrance of features of
the patients friend, seems an amazing coincidence that
demands explanation. The best explanation seems to
be that these elements of the dream have a common
cause, and that cause has to do with a thought about
the friend in question. (Glymour, 1983, p. 68)
Here Glymour observes that the fact that there is a connection in content between two dierent elements of the dream

1
Freuds conclusion after considering a number of examples of dreams that at rst sight appear not to be wish-fullments is as follows: I hope the
foregoing examples will be enough (till the next objection is raised) to make it seem plausible that even dreams with a distressing content are to be
construed as wish-fullments (Freud, 1991, p. 244). With this comment he seems to reject the obvious move that he could have made here, namely, to
claim that he now has enough inductive support to draw the general conclusion that all dreams are wish-fullments. Instead he cautiously claims only that
the examples make his claim plausible. He is not condent he has justied his theory. This is reinforced by a footnote he makes to the above sentence: I
must point out that the subject is not yet nally disposed of; I shall return to it later on (ibid.). This later on is most likely a reference to his longer
explanation of p. 709. There he asserts: Analysis is able to demonstrate that. . . unpleasurable dreams are wish-fullments no less than the rest. He goes on
to explicate this assertion with an account of the psychical processes involved in unpleasurable dreams that draws heavily on his theory of mind. It appears
that, at least in his more cautious moments, he holds his theory of dreams to be fully justied only with the support of his theory of mind.
2
Neil Manson (2003) argues that there has been a neglect by philosophers who defend psychoanalysis (Davidson, 1982; Hopkins, 1988; Wollheim, 1991;
Gardner, 1993; Cavell, 1993) of the role of Freuds theorising about the mind. He points out both the epistemological problems this creates for them and
the bizarre history of psychoanalysis it leads to.
3
While it is true that the woman had been free associating with the dream, this does not imply that each and every element of the dream played a causal
role in the thoughts so associated.

M. Michael / Stud. Hist. Phil. Biol. & Biomed. Sci. 39 (2008) 5264

and a thought that emerged through free association is

something that needs to be explained. Freuds theory explains this fact and appears to be the best explanation of
it. So Freuds method can be seen as involving an inference
to the best explanation. Even if one does not believe that
Freuds theory is the best explanation, the method can still
be seen as involving an attempt at an inference to the best
explanation. This is very dierent from asserting that the
only inference involved is a fallacious one based on an arbitrary transposition of a causal relation. Yet Glymour does
not acknowledge this. He continues to maintain that
Freuds method is fallacious in the structure of its logic,
even though there is an alternative and legitimate interpretation of the structure of its logic.
Glymour rapidly proceeds from his observation that
Freuds theory appears to be the best explanation of a surprising coincidence to an argument intended to undermine
the basis of this appearance. His argument is that the coincidence Freuds interpretation appears to explain is in fact
a manufactured coincidence:
one associates, at Freuds direction, until one thinks of
something which has connections with several elements
in ones dream; the several elements cause the common
thought, not vice versa, and the coincidence requires
no further explanation. (Ibid.)
If, as I argued previously, there is no obvious reason for
claiming the smoked salmon element as a cause of the
thought of the skinny friend, here we see a non-obvious
reason for doing so. This is that there is a fact that demands explanation: the coincidence that the thought of
the skinny friend connects two distinct elements of the
dream. The claim that these two elements were joint causes
of the thought explains that coincidence, in which case it is
in competition with Freuds explanation (that the elements
of the dream have a common cause).
If Glymour is simply assuming that the woman in the
example consciously associated with both the supper-party
and the smoked salmon element, then this, as I suggested
earlier, is a very uncharitable interpretation of Freuds
account. There is no indication, from what Freud says, that
the woman was trying to think of something which has
connections with all the elements in her dream. There is
no indication that she even regarded the smoked salmon
element of the dream as important, or had it on her mind
while she was free-associating.4 And there is no indication
that Freud prompted her to connect her associations in any
way with the smoked salmon element. Indeed, the smoked
salmon connection appears to be a surprise to the woman,
for she makes no mention of it while describing the situation regarding her skinny friend until she is reminded of
it by Freud. So if Glymours point is that the coincidence

55

is a result of a conscious attempt on the part of the patient

to make an association with connections to all the elements
of the dream, then this is an uncharitable presumption that
we may reasonably doubt. If, however, we (charitably)
interpret Glymours point as being that the best explanation of the coincidence is that the woman consciously associated with the smoked salmon element, then that is
something we should take more seriously. However, the
evidence from Freuds account still strongly suggests that
this was not the case. For why would the woman rst make
a number of associations that have no connection with the
smoked salmon element, and then, without mentioning it,
consciously nd an association that has a connection with
this element? This is unlikely to have been the result of
Freuds prompting, for he prompted for an association
with the smoked salmon element only after she had
described the situation regarding her skinny friend.
Perhaps though the smoked salmon element could have
had an unconscious causal inuence in bringing about the
thought of the skinny friend. If this is so, then the only evidence for such an unconscious causal inuence is the fact
that it explains the coincidence. So in order to explain a
coincidence, Glymour must claim that an element in memory can have an unconscious causal inuence on subsequent thoughts. But admitting this makes Freuds
explanation, and indeed his whole approach, inherently
plausible, for all he is saying in the context of the explanatory problem in question is that a thought which had
recently been on the patients mind had an unconscious
causal inuence on the content of the dream. There is no
dierence in initial plausibility between the two competing
explanations. So we may ask, in a competition between the
explanation that the two elements had as a common cause
the thought of the skinny friend, and the explanation that
the thought of the skinny friend was jointly caused by the
two elements, why favour the latter explanation?
Another interpretation of Glymours argument is that
the reason that the coincidence is manufactured is because
Freud selected from the many thoughts associated with the
dream only one that has connections with two or more elements of the dream. Thus the several elements cause the
common thought by constituting Freuds selection criteria.
And for this explanation we do not need to invoke unconscious causal inuences. But this explanation is incongruent with the account of the situation. There is no
indication in the account that Freud, in pinning his theory
on the thought of the skinny friend, had selected this
because of the smoked salmon connection. Indeed, he
seemed to have been just as unaware of the smoked salmon
connection at the time he formulated the hypothesis (that it
was thoughts about the skinny friend that caused the
dream content) as the woman was. So the proposition that

4
There is indication that she did regard the supper-party element of the dream as important and had it on her mind while she was free-associating. This
comes from her earlier associations, in which she talks of her husband as proposing to go on a strict diet and accepting no invitations to supper.

56

M. Michael / Stud. Hist. Phil. Biol. & Biomed. Sci. 39 (2008) 5264

the thought in question was caused by the smoked salmon

element by being selected by Freud for that connection is
not supported by the account.
There is a yet more decisive consideration that counts
against Glymours argument whatever interpretation we
put on it. Even if the woman had been consciously or
unconsciously looking for connections with both the supper-party and the smoked salmon element, it is surely a
remarkable coincidence she was so successful. That is, it
is a remarkable coincidence that there existed an association at all that had such striking connections with two distinctive elements of the dream. Moreover, it is even more of
a remarkable coincidence that the association in question
was one that had such emotional signicance for the
woman. The probability that Glymours explanation is true
can therefore only be as high as the expectation before the
connection was uncovered that there would be an emotionally charged thought that connects the two elements. But
the probability of Freuds explanation being true does
not depend on this expectation. Given this, Freuds explanation seems much more probable than Glymours.
There are additional considerations that greatly increase
the probability that Freuds explanation is the correct one,
ones that Glymour himself acknowledges:
The real coincidence is that on the one hand, on one day
the dreamer was visited by a friend, whose favourite dish
is smoked salmon, and who had asked her when she, the
dreamer, planned to have another supper-party, and, on
the other hand, that night she dreamed of a supper-party
and smoked salmon. Is this coincidence evidence of a
causal connection between the encounter with her
skinny friend in the dream? Perhaps it is, in the case at
hand, for the real event and the dream are proximate
in time and share a number of independent features,
and one doubts that any event so proximate would share
those features. (Glymour, 1983, p. 69)
This is a very grudging admission. When one takes into account the event in question, the proximity to the dream,
and the number of independent features that the event
and the dream share, Freuds explanation appears far more
probable than any alternatives. Also to say this is the real
coincidence and not the earlier one is of little consequence,
for Freuds argument has always been that there is a mesh
of coincidences that his interpretations explain, hence the
explanation of each coincidence is supported by the explanation of each of the others.
Even the above admission is not enough to persuade
Glymour of the validity of Freuds method:
Freud is not so much attempting to convince us of the
correct interpretation of a specimen dream as to convince as of a method, and, once again, there is no reason
to think the method trustworthy in general. As the span
of time between the real event and the dream increases
and in many of Freuds applications it increases to
yearsthe chances of coincidence increase immeasur-

ably, and the inference to a causal connection becomes

increasingly ill founded. (Ibid.)
All the indications are that Freuds interpretation in this instance is correct, yet Glymour still does not think that this
is favourable for Freuds method. True, the devil is in the
detail, but surely success oers positive support for the
method. What Glymour should have concluded from his
analysis of the dream in question is that Freuds method relies, not on an unjustied transposition of a causal relation,
but on inferences to the best explanation. The coincidence
that impressed Glymour in this dream interpretation was
the proximity of the event that engendered the explanatory
dream-thought. What is to say that for another dream
interpretation it would not be some other equally impressive coincidence that clinches the case? One must consider
each example that Freud presents on its own merits (moreover, if Glymour had another, unfavourable, example in
mind, he should have presented that instead!). What matters here is that Freuds method, as demonstrated by this
particular example, has yielded a successful interpretation,
and we see its success as being based on an inference to the
best explanation, which is a ubiquitous and valid form of
inference.
3. Cashing out the metaphors
The above analysis of Glymours argument oers some
hints as to how to cash out the metaphors of the crossword
and jigsaw puzzles, and hence how to reply to the many
interpretations objection and to justify Freuds dream
interpretations. As we saw in the above example, the devil
is in the detail, and no general argument can validate all, or
indeed any, of Freuds interpretations. Each interpretation
is an inference to the best explanation built up from smaller
inferences to the best explanation. But in each such inference, whether what is inferred is the best explanation or
not may come down to a ne judgement that depends on
the particular explanandum in question. So a formal argument with general scope can only hope to show how, if the
method is applied correctly, the interpretation may be
valid. It cannot prove the validity of any particular
interpretations.
Dreams raise explanatory problems. The main one is
this: Why did I dream that particular dream? This general
problem can be decomposed into smaller problems: Why
did that particular detail turn up in my dream? Each
dream therefore presents a series of explanatory problems
and so the best explanations of these would, if good
enough, be acceptable conclusions of inferences to the best
explanation.
Consider the case of a dreamer who is puzzled by two
dierent elements of her dream, A and B. Suppose that
the dreamer associates with A a particular memory M
(where M is perhaps a concern of the dreamer, a remembered event, a thought from the previous day, and so
on), such that there is a connection in content between A

M. Michael / Stud. Hist. Phil. Biol. & Biomed. Sci. 39 (2008) 5264

and M. We can draw the explanatory hypothesis that A

was in the dream because M had been on the dreamers
mind. Suppose further that soon afterwards it is recognised
that M also oers an easy solution to the other puzzling
element: the dreamer nds a connection between M and
B. The hypothesis that A was in the dream because M
had been on the dreamers mind is now more probable.
This is the essence of the crossword analogy: if we nd
what we think might be a solution to a clue, and that solution provides a letter that immediately suggests a solution
to another clue, then that increases the probability that that
solution is correct. The reason why the hypothesis is more
probable is that now there are more facts that it explains.
The hypothesis is that, (i) M was the cause of the presence
of A in the dream, which implies that, (ii) M was causally
active at the time of the dream. The hypothesis now plays
a role in explaining three things:
1. (i) explains the presence of A in the dream,
2. (ii) increases the probability of the following: (iii) M
was the cause of the presence of B in the dream, which
explains the presence of B in the dream,
3. (i) and (iii) together explain why M has a connection
with both A and B.
Notice, incidentally, that the hypothesis may be a good
explanation of the presence of A even if there had been no
connection with B or any other element of the dream. This
would be the case if there were independent evidence to
suppose that M may have been on the dreamers mind at
the time of the dream (that is, was recently causally active
in generating thoughts), for example, if M had been part of
the thoughts and concerns of the previous day. So the
hypothesis may already have a great deal in its favour, even
before we consider other elements of the dream it might
explain.
There is however an alternative to Freuds solution to
the explanatory problems raised by dreams, and that is that
the elements of the dream are generated randomly (dreams
are froth). A dream, on this view, is just an arbitrary
quasi-narrative concoction from various snippets of memory, so there is no particular reason why the dreamer
dreamt what he did. Does this oer a better explanation
of why the dreamer dreamt of A and B in the above example? I think that intuitively the rst explanation is the better
one. This may be because it appeals to an underlying commitment to the idea of order in the world, including order
in the mind, and explanations that invoke randomness are
intuitively less satisfactory than ones that invoke order are.
Also, we already know that there are many dreams in
which the content is not generated randomly, but has various connections, for example, with what happened the day
before or with ones wishes (as in simple thirst, hunger, and

57

sex dreams). This gives us yet more reason to prefer an

explanation that does not invoke randomness. More pertinently though, if we accept the randomness explanation
then we still leave one fact unexplained: how is it that the
dreamer freely associated with A an element M that happened to have a connection with both A and B? This seems
unlikely to happen by chance alone.
I will try to make this argument more formal. Consider
the subjects memory as a set of self-contained snippets,
which may be words, phrases, images, events, and so on.
Imagine picking out an item, A, from this set at random.
Dene the set S1(A) to be the set of every item in memory
that has an immediate connection with A, the set S2(A) to
be the set of every item in memory that has an immediate
connection with an item in S1(A), and so on. A connection
between two elements may take many dierent forms:
semantic, phonetic, historical, and so on. How striking or
tenuous the connection is will be a matter of degree. A connection may be intersubjectively striking, in the sense that
anyone would see it as a striking connection, or it may
be subjectively striking. A connection would be subjectively
striking if it is personally meaningful to the dreamer, that is,
if it has some kind of personal resonance for her. It is subjective in the sense that what makes it striking is the degree
of personal resonance the connection has for the dreamer
rather than someone else, not in the sense that it cannot
be independently veried to have such personal resonance
for the dreamer based on what is known about her.5 In
any case, the dreamers judgement of the quality of the connection is a necessary determinant of whether or not the
connection is relevant to the interpretation. Given striking
enough connections, I think that S1(A) will be a very small
subset of the subjects memory.
Suppose that in a given dream interpretation the rst
item, M, associated by the subject that has a connection
with an element of the dream, A, happened to also have
a connection with another element, B. Then M 2 S1(A) \
S1(B). It is likely that S1(A) \ S1(B) is a small subset of
S1(A), so the probability of picking an item at random from
S1(A) and nding that that item is also in S1(B) is likely to
be small. Its value is given by jS1(A) \ S1(B)j/jS1(A)j (where
jXj is the number of elements in the set X). The probability
that it was a pure coincidence that the dreamer just happened to associate with A an item that had a connection
with B cannot be more than jS1(A) \ S1(B)j/jS1(A)j. The
more striking the connections between M and B, the smaller this probability will be. The smaller this probability, the
smaller the probability that the randomness solution is
correct.
There are a number of objections that could be made to
the above argument. The rst is that during free association, more than one sampling of the set S1(A) is made, so
the probability that the dreamer would pick out an element

5
It will be important from an epistemological point of view that a subjectively striking connection should be one that would have been considered
striking by the dreamer over a reasonable period of time, certainly starting from before the dream, and not just during the interpretative process.

58

M. Michael / Stud. Hist. Phil. Biol. & Biomed. Sci. 39 (2008) 5264

that happened to also lie in S1(B) is actually higher than

jS1(A) \ S1(B)j/jS1(A)j. This is true, but if S1(A) is a relatively large set and S1(A) \ S1(B) is a relatively small set,
and if the number of samplings of S1(A) are small compared to jS1(A)j, then the probability of picking an item
from S1(A) \ S1(B) still remains small. A second objection
is that there is actually a selection bias introduced by the
interpreter, in that the dreamer will be encouraged to associate with A until she comes up with an item that has a connection with B. So then the probability of picking out an
item from S1(A) \ S1(B) is not limited. There are a number
of things that count against this objection. First, it is a matter of empirical observation whether this kind of selection
bias actually takes place. As we saw in our discussion of
Glymour, in the example given there was no evidence of
such selection bias. Secondly, if S1(A) is a relatively large
set and S1(A) \ S1(B) is a very small set, then the process
may go on for a long time before an appropriate item is
found. Again, this does not seem to t in with the facts
about Freuds actual interpretations, in which the duration
of free association is not usually very long.
There is however an alternative explanation for why the
dreamer associated with A an element that happens to have
a connection with both A and B. This is that the person
who had the dream, when asked to associate with A,
already had on her mind B as well, and that this had an
unconscious causal inuence on her choosing M to associate with A rather than any of the other things she could
have associated with it. If this were the case, then the fact
that M is also connected with B may not signicantly
decrease the probability of the randomness solution.
However, it would be ironic that in order to defend the
randomness explanation against one which favours
unconscious inuences and underlying mental order one
needs to invoke unconscious inuences and underlying
mental order. If it is plausible that having elements A
and B on her mind unconsciously inuenced the dreamer
to associate with A an element M that has an immediate
connection also with B, then it is no less plausible that having element M on her mind unconsciously inuenced her to
dream of A and B. In other words, the higher the initial
plausibility of the unconscious inuence explanation, the
higher the initial plausibility of the Freudian explanation.
There is a yet more decisive consideration that counts
against both the selection bias and the unconscious inuence objections. Even though there may not be a limit to
the probability of picking out an item from S1(A) \ S1(B),
there is still a limit to the probability of the randomness
explanation being true. This is the expectation that
S1(A) \ S1(B) 6 . That is, the probability that the randomness explanation is true is limited by the probability
that given two random items such as A and B, we would
have S1(A) \ S1(B) 6 . Given that S1(A) and S1(B) are
likely to be small subsets of memory, this probability is

See Hobson (1988) and Porte (1988).

likely to be small. In other words, the very fact that there

existed an item in memory that could be connected with
both A and B is a coincidence whose prior probability is
likely to have been low. Again, this depends on just how
striking the connections between M and A and B are.
Moreover, if we require beforehand that the only associations relevant to an interpretation are ones that are emotionally signicant to the dreamer, then the expectation
that there exists an appropriately emotionally signicant
item in S1(A) \ S1(B) is even smaller. The likelihood of
the randomness solution becomes even lower the more elements of the dream that are unied by an association. If we
have three randomly selected items, A, B, and C, the expectation that S1(A) \ S1(B) \ S1(C) 6 is even smaller than
the expectation that S1(A) \ S1(B) 6 .
Everything that has been said so far against the randomness explanation is also relevant against other explanations, such as that of the somatic inuence on dream
content.6 For the Freudian explanation, as we have seen,
explains not only dream content but also the various connections that emerge during the association process, and
from the discussion above we have seen how alternative
explanations of these connections, such as selection bias
and unconscious inuence, may not be as good as the
Freudian explanation.
This analysis begins to cash out the metaphors Freud
uses to defend his method. The conclusion is that if Freud
applied his method rigorously, then his interpretations may
be justied as inferences to the best explanation. A rigorous
application of the method will be one where the associations are drawn freely by the dreamer and not suggested
by the analyst. Given this, there are a number of factors
that increase the epistemological validity of the interpretation: (i) how striking the connections are: the more striking,
the more likely the interpretation; (ii) the number of elements unied by a particular association: the more elements unied, the more likely the interpretation; (iii) the
number of links in the chain of associations: the smaller
the chain the more likely the interpretation.
4. The dream of Ottos illness
To make this discussion more concrete I will look at a
particular criticism of Freuds interpretations. Rosemary
Sand describes two approaches to dream interpretation,
Freuds and the classical method Sand (1993). An example of a classical dream interpreter is Artemidorus of Dalsis, from the second century (who, incidentally, Freud
mentions on several occasions as one of his inspirations),
who believed that it is natural for a lover to seem to be
with his beloved in a dream and for a frightened man to
see what he fears. This kind of interpretation is about
making simple connections between the manifest content
of a dream and the hopes, fears, desires, and other concerns

M. Michael / Stud. Hist. Phil. Biol. & Biomed. Sci. 39 (2008) 5264

of the dreamer. So a dream that represents love is about

love, one that represents fear is about fear, and so on. This
is distinguished from Freuds method by the absence of the
use of free associations. It is not clear, from what Sand
says, how banal an interpretation needs to be in order to
qualify as a classical interpretation, but it seems any remotely surprising interpretation ought to be ruled out, which
rules out most of Freuds. Sand argues in favour of the
classical method of interpretation and against Freuds
method. Her criticism of Freud is that he makes the free
association fallacy: given a dream A, if I associate from
A to B, then B must have been a determinant of A. To
illustrate the fallacy Sand considers Freuds interpretation
of one of his own dreams. This is the dream:
My friend Otto was looking ill. His face was brown and he
had protruding eyes.
Otto is my family doctor, and I owe him more than I can
ever hope to repay . . . He had visited us on the dreamday, and my wife had remarked that he looked tired
and strained. That night I had my dream, which showed
him with some of the signs of Basedows disease. (Freud,
1900/1991, p. 372)
Sand makes the following comments:
Freud then free associated to this dream and after a long
series of thoughts arrived at what he considered to be
the latent dream, two wishes, one that he might become
a professor and the other that he might live a long life. . .
The assumption that the supposed latent dream, the
wishes for a professorship and a long life, gave rise to
the manifest dream about Ottos appearance is based
solely on the fact that a string of associations connected
them. . . This conclusion seems entirely unjustied unless
it is simply presupposed that ideas which emerge in association to a dream preceded it and produced it.
I do not think that this is a fair description of the logic
of Freuds argument. What it misses is the explanatory role
played by Freuds interpretation. It is not simply presupposed that ideas associated are determinants. Rather, that
the ideas associated are determinants is oered as the best
explanation of certain puzzling features of the dream.
Freud specically makes reference to such a feature:
Anyone who interprets this dream without regard for
my rules will conclude that I was worried about my
friends health and that this worry was realized in the
dream. . . But I should be glad if anyone interpreting
the dream in this way would be good enough to explain
to me why my fears on Ottos behalf should have lighted
on Basedows diseasea diagnosis for which his actual
appearance gives not the slightest ground. (Freud,
1900/1991, p. 372)
The interpretation that Freud considers and dismisses here
is exactly the kind of interpretation that would be given by
the classical method that Sand describes. Freud explicitly
states what he nds unsatisfactory about this: it leaves a

59

prominent feature of the dream unexplained. This observation is pertinent to all those critics who analyse Freuds
method as relying on the free association fallacy, for it
clearly brings out the explanatory nature of Freuds
method.
By this stage, Freud already has rules to guide him in
the interpretation of dreams, namely, to only consider egoistic impulses as possible causes of the dream and to look
for a wish that is fullled in the dream. Nevertheless, even
before he makes use of these rules, he uncovers material
that goes deeper than the classical method could:
My analysis. . . brought up the following material from
an occurrence six years earlier. A small group of us,
which included Professor R., were driving in pitch darkness through the forest of N., which lay some hours
drive from the place at which we were spending our
summer holidays. The coachman, who was not perfectly
sober, spilt us, carriage and all, over an embankment,
and it was only by a piece of luck that we all escaped
injury. We were obliged, however, to spend the night
in a neighbouring inn, at which the news of our accident
brought us a lot of sympathy. A gentleman, with unmistakable signs of Basedows diseaseincidentally, just as
in the dream, only the brown discoloration of the skin of
the face and the protruding eyes, but no goitreplaced
himself entirely at our disposal and asked what he could
do for us. Professor R. replied in his decisive manner:
Nothing except to lend me a night-shirt. To which
the ne gentleman rejoined: Im sorry, but I cant do
that, and left the room.
As I continued my analysis, it occurred to me that Basedow was the name not only of a physician but also of a
famous educationalist. . . But my friend Otto was the
person whom I had asked to watch over my childrens
physical education, especially at the age of puberty
(hence the night-shirt), in case anything happened to
me. (Ibid., pp. 372373)
Thus far Freud has associated two items with (A) Basedows disease: one is (M1) the memory of an event that
took place six years ago, the other is (M2) the memory that
there is a famous educationalist called Basedow. With the
latter, Freud associates (B) his friend Otto who was in
the dream, via the following thought (T): Otto was (t1)
the person asked to watch over his childrens physical education in case anything happened to him, especially (t2) at
the age of puberty, which he associates with a night-shirt,
an element from M1.
Freud has, from the material uncovered, a means for
partially explaining the presence in the dream of the puzzling element A, namely, that M1 was on his mind when
he had the dream and produced A. What are the alternatives to this explanation? A classical interpretation along
the lines described by Sand oers no explanation, as Freud
has shown. One way of supplementing this interpretation,
however, would be to invoke randomness: Basedows disease was an element from Freuds memory that randomly

60

M. Michael / Stud. Hist. Phil. Biol. & Biomed. Sci. 39 (2008) 5264

turned up in his dream. The judgement then is between an

explanation that states that M1 was the cause of A and an
explanation that states that A was generated randomly.
I will rst concentrate on something that Freud mentions almost in passing, and so is easily missed: the distinct
similarity between the appearance of the diseased man in
M1 and Otto in the dream (just as in the dream, only
the brown discoloration of the skin of the face and the protruding eyes, but no goitre). I think that it is possible to
infer that M1 was the source of the dream-image on the
grounds of this connection alone. This would be an inference to the best explanation. How strong the inference is
depends on how distinctive the dream-image is, and only
Freud could answer that question. This qualication notwithstanding, my argument is as follows. The elements of
the dream content must come from the dreamers memory.
The issue then is which items of memory they come from.
Consider the image in Freuds dream of Otto, having the
unmistakable signs of Basedow diseasebrown discoloration and protruding eyesbut, unusually, no goitre. It
is possible that such an image was put together from
numerous smaller elements of memory, but I think that if
it could be shown that there exists a self-contained image
in memory which matches it as a whole, then the more
probable explanation is that the image in the dream comes
directly from this self-contained image, for this is the more
parsimonious explanation. Similarly, the image may have
come from a deduction of elements from a more detailed
image, but if it could be shown that there exists an image
with the same amount of detail in memory, then it is more
probable that the dream-image came from this, for, again,
this is the more parsimonious explanation. Now there is a
relatively self-contained image in memory that matches the
dream-image, and that is the one from M1 (the only dierence is that Otto replaces the gentleman of that incident). If
the dream-image is highly distinctive then it is unlikely that
there is another such self-contained image in memory,
hence it is most probable that M1 is the source of the
image.
However, the situation is even more favourable for this
conclusion, as there is more than just the one connection
that emerged during Freuds period of free association.
As I argued in the previous sections, where there are a
number of connections between the dream and items associated with the dream, the existence of these connections
must be explained. Fig. 3 is a diagram outlining all the connections that emerged in the analysis of the dream of Ottos
illness. The judgement we need to make is this: is it more
likely that A was randomly generated and that Freud
would freely associate with A an item M1 that just happened to have the above connections, or is it more likely
that M1 had a causal inuence on the dream content? If
one judges that the former is less likely than the latter,
and there are no other explanations on oer, then the latter
can be inferred to as the best explanation of A.
What I have said so far is a sucient reply to Sands
argument about the free association fallacy. The argu-

ment that M1 preceded and produced the dream element

A, is that this is the best explanation of two facts: the presence of A in the dream and the connections uncovered during free association. So it is not a fallacy to argue that an
item freely associated with an element of the dream preceded and produced that element, for such an argument
is, as we have seen, an inference to the best explanation.
One can dispute whether or not it really is the best explanation, but this would be an argument over probabilities,
which would depend on some ne judgements. What matters is that the conclusion reached by such an argument is
not entirely unjustied, nor does it simply presuppose
anything. Sand and the other advocates of the free association fallacy are wrong on these points.
However, the conclusion reached thus far, that the presence of A was due to M1 being on Freuds mind at the time
of the dream, is only a partial explanation for the presence
of A, for now we have a further explanatory problem: why
was M1 on Freuds mind at the time of the dream? We cannot regard the presence of A as satisfactorily explained
unless we have an answer to this question, since the problem of why M1 was on Freuds mind at the time of the
dream is the same type of explanatory problem as that of
why A was in the dream.
Freud proceeds to give the following interpretation of
his dream:
By giving my friend Otto in the dream the symptoms of
our noble helper, I was evidently saying that if anything
happened to me he would do just as little for the children
as Baron L. had done on that occasion in spite of his kind
oers of assistance. This seems to be sucient evidence
of the egoistic lining of the dream. (Ibid., p. 373)
Although he is guided by the rules of his method in coming up with this interpretation, the interpretation can be
considered as an explanation of why he had M1 on his
mind at the time of the dream in its own rights. This explanation links M1 with certain deeper feelings and concerns
that Freud attributes to himself. The rst part of the interpretation works on a parallel established by positing M1
and M2 to be etiologically associated with the contents of
the dream. The dream association between Otto and Basedows disease and the connection between A and Baron L.
(the gentlemen with Basedows disease in M1) establish a
connection of identication between Otto and Baron L.,
in that they are represented as both having Basedows disease and as having the same appearance. The connection
between T and the request made in M1 establishes a connection between the request made to Baron L. by Professor
R. to provide him with a night-shirt and the request made
to Otto by Freud to look over his childrens physical education, especially at the age of puberty. Given the connection between the two requests and the connection between
the people the requests were made to, it is plausible that
there are further connections closely relating to these two
situations. Freud posits that there is such a connection
between the answer given by Baron L. and the answer he

M. Michael / Stud. Hist. Phil. Biol. & Biomed. Sci. 39 (2008) 5264

A
a1

61

B
a2

m21 m22
M2

T
t1

t2

m11 m12 m13

M1
KEY:
A

dream element: looking ill, brown discoloration and protruding eyes

a1

Basedows disease

a2

only brown discoloration and protruding eyes, no goitre

dream element: Otto

thought: had asked Otto to watch over childrens education

t1

the person whom I had asked to watch over my childrens physical education. . .

t2

. . . especially at the age of puberty

M1

memory: carriage-spilling event of six years earlier

m11

gentleman with unmistakable signs of Basedows disease

m12

only brown discoloration and protruding eyes, no goitre

m13

night-shirt

M2

memory: Basedow the educationalist

m21

Basedow

m22

education

Fig. 3. Outline of the initial connections that emerged in Freuds analysis of his dream of Ottos illness.

feared would eectively be given in the case of Otto and his

children. Though Freud may not have been conscious of
such a fear before his self-analysis, it is highly plausible that
he would have had such a fear given the events of the previous day (his wife observing how tired and strained Otto
had looked). By positing such a feeling and such a connection Freud was oering an explanation of why M1 was on
his mind by linking those aspects of M1 already linked with
the contents of the dream with a deeper feeling and concern
of his.
It is a perfectly natural move to try to explain why we
have a certain thought on our minds by seeking some deeper feeling or concern as the cause. This is a regular and

unremarkable feature of the common-sense explanation

of our thoughts, which occurs, for example, when we have
lost the thread of our thoughts and to try to recapture it we
look for what would have been the starting point. As an
explanation of why Freud had M1 on his mind, the
account given by Freud faces two competitors. The rst
is the complete randomness explanation, that M1 just
randomly happened to be on his mind. If we have already
accepted that positing M1 oers a better explanation of the
presence of A in Freuds dream than randomness, then it is
unlikely that we would now see randomness as a better
explanation of the presence of M1 on Freuds mind than
an account like the one Freud gives. For though we may

62

M. Michael / Stud. Hist. Phil. Biol. & Biomed. Sci. 39 (2008) 5264

have accepted the etiological presence of M1 as the better

explanation on the grounds that it explains the connections
that emerged subsequently, having accepted it, the initial
plausibility of a randomness solution is reduced for us.
Moreover, the Freudian account better explains the connections between M1 and M2 than if M1 and M2 had been
generated at random. The second competitor for Freuds
account is an explanation that allows for a chain of associations but does not require the linking of these with any
deeper feeling or concern Freud may have had. For example, with Otto on his mind the day before, Freud may have
made the following circular chain of associations that
brought up both M1 and M2: Otto ? look after childrens
education at puberty ? night-shirt ? Baron L. ? M1 ?
Basedow ? M2 ? educationalist ? Otto looking after
children at puberty. However, what Freuds account
explains better is why those particular associations were
made. Many things could have been associated with each
item in the chain, but why was it those particular associations that were drawn? Again, randomness may be a solution, but given the emotional poignancy of one of the items
in the chain (the one involving Freuds children) and our
natural tendency to look for feelings and concerns in
explaining our threads of thought in common-sense psychology, Freuds explanation is surely preferable.
We now have the connections shown in Figure 4.
Freuds concern (C) that Otto would not be able to look
after his childrens education if anything happened to
him explains why M1 was on his mind at the time of the
dream. It also increases the probability of the hypothesis
that M1 was on his mind at the time of the dream and
caused A, since there is now another connection that would
need explaining by those advocating randomness, namely,
that between the request and answer of M1 and the request
and anticipated answer of Freuds concern.
Freud now has an interpretation of the dream that links
its manifest content to a concern of his via a network of
connections. The concern uncovered, Freuds fear that
Otto would not be able to look after the education of his
children if anything happened to him was one that may
have remained hidden had he not proceeded to interpret
his dream. Moreover, it is a surprising interpretation, and
would never have been gleaned from the manifest content
of the dream had Freud not employed his method of free
association. It is clearly therefore not a classical interpretation along the lines Sand describes. It is, if one accepts the
judgements made along the way, the best explanation of
why the dream had the manifest contents it did. As such,
we have both refuted the so-called free association fallacy
and shown how a Freudian interpretation may be justied.
We have also shown how a conclusion may be logically
compelled by the internal explanatory dynamics of the
interpretation.7

M2

M1

Fig. 4. General outline of the connections between elements in the

contents of the dream of Ottos illness and the associations Freud made
with these.

What I have said so far vindicates Freuds method of

dream interpretation, but not his theory of dreams. For
there remains the small detail of showing how this dream
is a wish-fullment (and the even smaller detail of showing
that it is a fullment of an infantile wish, but even Freud
does not bother about that). It is in this that we nd an
epistemological weakness in Freuds interpretation:
But where was its wish-fullment to be found?. . . At the
same time as I represented Otto in the dream as Baron
L., I had identied myself with someone else, namely
Professor R.; for just as in the anecdote R. had made
a request to Baron L., so I had made a request to Otto.
And that is the point. Professor R., with whom I should
really not venture to compare myself in the ordinary
way, resembled me in having followed an independent
path outside the academic world and had only achieved
his well-merited title late in life. So once again I was
wanting to be a Professor! Indeed the words late in life
were themselves a wish-fullment; for they implied that I
should live long enough to see my boys through the age
of puberty myself. (Ibid., p. 374)
All of this is immensely plausible. The problem is, in
order to justify it as we have done in the previous cases,
we need to show that it is an inference to the best explanation. But what exactly does it explain? Positing that M1
was on Freuds mind at the time of the dream served to
explain the puzzling presence of Basedows disease in the
dream. Positing Freuds fear about Ottos inadequacy
served to explain why M1 was on Freuds mind at the time
of the dream. But now there is nothing left to explain. If it
does not explain anything, then however plausible it is,
however neatly it ts in with the other parts of the

7
I of course do not mean logically compelled in a deductive sense here, but only in the sense in which one conclusion stands out as better supported by
the evidence than any other.

M. Michael / Stud. Hist. Phil. Biol. & Biomed. Sci. 39 (2008) 5264

interpretation, it cannot be justied as an inference to the

best explanation. And if it is not justied as an inference
to the best explanation, then it is dicult to see how it
could be justied at all. If Occams razor had been applied
to Freuds interpretation, then this piece would have been
shaved away.
5. Conclusion
In this essay I have shown that two of the main arguments brought against the validity of Freuds dream interpretations do not stand up to criticism. The free association
fallacy misconstrues the logic of Freuds method: it does
not involve an unjustied reversal of causal relations, but
rather a legitimate attempt at an inference to the best
explanation. Though it is true that the dream determines
the material that comes up during free association, this fact
does not play an important role in the inference. What the
inference crucially explains is thematic connections between
the material uncovered and the dream, not causal relations.
Freuds interpretations solve two explanatory problems:
(1) why did the dream contain those elements, and (2) why
does the material uncovered during free association have
numerous connections with dierent elements of the
dream. It is in also solving (2) that the Freudian explanation of (1) is superior to other explanations. For it is a surprising coincidence that an association has striking
connections with two or more dierent elements of the
dream. Indeed, it is a surprising coincidence that any item
has striking connections with two or more dierent elements of the dream. The more striking the connections,
the bigger the coincidence, the less likely that this happened
by chance. Freuds interpretations explain this coincidence
by positing that the item in question had a causal inuence
on the content of the dream. I have considered a number of
alternatives to this explanation, based on an argument by
Glymour, and have shown that when considered in the
context of an example (one that he chose) they are inferior
to the Freudian explanation.
Three additional points need to be made. One is that,
even if there was an explanation of (2) that was as good
as the Freudian explanation, Freuds explanation still has
the advantage over other explanations of (1) in that it
explains (1) and (2) simultaneously. A unied theory that
explains two facts simultaneously is preferable to one that
explains only one fact, as long as there is nothing else to
choose between those two interpretations. The second
point is that, though chance may explain the coincidence
in (2) on one occasion, the more we nd such coincidences,
the less plausible the chance explanation becomes. And this
is the case with Freudian interpretations, where it is a matter of course that during free associations material emerges
with connections to two or more dierent elements of the
dream. The third point is that, if any particular Freudian
interpretation can be proved to be true, then this gives us
more reason to believe that any other Freudian interpretation is also true.

63

Though the judgement of whether or not a particular

Freudian interpretation is the best explanation depends
on the details of the case in question, there is no general
reason to suppose that it cannot be, and there are general
reasons for supposing that, if the method has been applied
rigorously, then it might be. I have tried to formalise these
reasons by providing an outline of the kinds of probabilities that would be involved in making a judgement between
the Freudian and the randomness explanations. This
account also helps to cash out the metaphors of the jigsaw
and crossword puzzles that Freudians have used in
response to the many interpretations objection. It shows
how the numerous connections uncovered during free association, along with other factors such as the emotional signicance of the associations, the distinctiveness of the
dream images, and the historical proximity of events,
may logically compel a certain interpretation. The evidential base here is so rich that it is unlikely that there are
numerous equally plausible explanations, so the many
interpretations problem does not arise. This is so as long
as what is inferred really does play a crucial role in solving
the explanatory problems posed. We saw in the case of the
dream of Ottos illness, where Freud had a well supported
and explanatorily exhaustive interpretation but went on,
due to his theoretical commitments, to add to this two
more explanatory items (the wishes), that there is a temptation to expand the explanation beyond what is required by
the explanatory problem. However, the only reason Freud
proceeded to make this expansion was his theoretical commitment to the hypothesis that all dreams are wish-fullments. So we must distinguish between how much of
Freuds interpretations are driven by the internal logic of
the interpretative process and how much are driven by
his theoretical commitments, which have more to do with
his general theorising about the mind. In this essay I have
concentrated solely on the former. As such, I think that
Freuds method of dream interpretation is, in its purest
form, both independent of his theory of dreams and logically valid.
References
Cavell, M. (1993). The psychoanalytic mind: From Freud to philosophy.
Cambridge, MA: Harvard University Press.
Davidson, D. (1982). Paradoxes of irrationality. In R. Wollheim, & J.
Hopkins (Eds.), Philosophical essays on Freud (pp. 289305). Cambridge: Cambridge University Press.
Freud, S. (1961). Remarks on the theory and practice of dreaminterpretation as a whole. In J. Strachey (Ed. and Trans.), The
standard edition of the complete psychological works of Sigmund
Freud (Vol. 19, pp. 109121). London: Hogarth Press. (First published
1923)
Freud, S. (1991). The interpretation of dreams. The Penguin Freud Library,
4. London: Penguin. (First published 1900)
Gardner, S. (1993). Irrationality and the philosophy of psychoanalysis.
Cambridge: Cambridge University Press.
Glymour, C. (1983). The theory of your dreams. In R. S. Cohen, & L.
Lauden (Eds.), Physics, philosophy and psychoanalysis (pp. 5771).
Boston Studies in the Philosophy of Science, 76. Dordrecht: D. Reidel
Publishing Company.

64

M. Michael / Stud. Hist. Phil. Biol. & Biomed. Sci. 39 (2008) 5264

Grunbaum, A. (1984). The foundations of psychoanalysis. Berkeley & Los

Angeles: University of California Press.
Hobson, J. (1988). Psychoanalytic dream theory: A critique based on
modern neurophysiology. In P. Clark, & C. Wright (Eds.), Mind,
psychoanalysis, and science. Oxford: Blackwell.
Hopkins, J. (1988). Epistemology and depth psychology: Critical notes
on The foundations of psychoanalysis. In P. Clark, & C. Wright
(Eds.), Mind psychoanalysis and science (pp. 3360). Oxford:
Blackwell.
Lipton, P. (1995). Prediction and prejudice. In idem (Ed.), Theory,
evidence and explanation. Aldershot: Dartmouth Publishing
Company.

Manson, N. (2003). Freuds own blend: Functional analysis, idiographic

explanations, and the extension of ordinary psychology. Proceedings of
the Aristotelian Society, 2, 179195.
Porte, H. (1988). The analogy of symptoms and dreams: Is Freuds dream
theory an impostor? In P. Clark, & C. Wright (Eds.), Mind,
psychoanalysis, and science (pp. 321342). Oxford: Blackwell.
Sand, R. (1993). On a contribution to a future scientic study of
dream interpretation. In J. Earman et al. (Eds.), Philosophical
problems of the internal and external worlds: Essays on the
philosophy of Adolf Grunbaum (pp. 527546). Pittsburgh: University
of Pittsburgh Press.
Wollheim, R. (1991). Freud (2nd ed.). London: Fontana Press.