Biting the Bullet

Fantasy and the Psychoanalytic View of Human Beings

The word 'fantasy' refers to ideas and streams of thoughts, often accompanied by emotions. On the one hand, we have to make a distinction between fantasies and 'normal' propositional attitudes (beliefs, desires, etc.): the former may be unrealistic or even impossible. On the other hand, there is a difference between fantasies and delusions (hallucinations, confabulations): fantasies are thought to be non-pathological as such-part of our normal life. However, in psychoanalysis, they are often 'dubbed', that is, labelled with epithets such as, 'narcissistic' or 'omnipotent', and there is also talk about 'pathological' fantasies.

Fantasies, silly jokes and weird ideas belong to our daily life, and novelists and screenwriters systemically make use of fantasies. Psychoanalysts lean on them, too, and most analysands know that. Given the central rule of the psychoanalytic cure (analysands are advised to say whatever comes to their mind, whether it appears relevant and important or not), the psychoanalytic method supports analysands talking about their fantasies. We could say that psychoanalysis in general contains a fantasy-like view of human beings: we are driven not only by logical or rational rules (the reality principle), but also by conflicting and non-rational urges (the pleasure principle). Thus, an analysand's mind is, for the most part, not studied through logical relations of matters, but through her/his (non-logical) 'free' associations. Thus, it is found that (apparently) distinct matters relate to and affect each other.

An analysand's ideas concerning her/his analyst (transference reactions), for example, often seem to come close to those concerning her/his parents, teachers and other authorities. Transference often includes fantasies (and 'primal fantasies'), which are developmentally based on early bodily sensations, and later formed into images, symbolic representations and even verbal expressions. Psychoanalysts claim to have access to the dynamic 'psychic reality' of fantasy. Past thoughts, attitudes and emotions are fantasized through transference phenomena 'as if' they were currently real, whether in the psychoanalytic or in the everyday situation. The changes of transference are seen to reflect 'ongoing' fantasizing.

Making the dynamics of an analysand's mind apparent in the talking cure is thought to contain curative power: in this way it is possible to solve psychic conflicts (or, at least, to make them more tolerable). It is an important characteristic of the psychoanalytic cure: instructions or advice are not given, but the analyst pays 'free-floating' attention to the (even) weird and irrational ideas and feelings of an analysand. From this perspective, psychoanalysis is the art of fantasy.

Fantasy in Psychoanalysis and Cognitive Science

We arrive at the core of the psychoanalytic conception of fantasy through Freud's idea that there are two principles in mental functioning: the pleasure principle and the reality principle (see, e.g., Freud, 1911Freud, /1953c. In infancy, thinking is characterized by seeking pleasure and (illusionary) wishfulfillment. The realities of the outer world are increasingly taken into account in the thought processes, and there is a gradual transition from the pleasure-ego to the reality-ego. However, the pleasure principle does not become wholly displaced by the reality principle. Even in adult mental life, thinking contains aspects that aim at seeking pleasure and the fulfillment of wishes regardless of the restrictions imposed by reality. This aspect of thinking is most clearly seen in dreams and fantasies. For example, unsatisfied wishes were seen by Freud (1908Freud ( /1953b as 'driving forces' ('Triebkräfte') of fantasy, and every fantasy was a wish-fulfillment, a 'correction of unsatisfactory reality'.

The English translators of Freud began to use the word 'phantasy', with the distinguishing ph, in order to refer to predominantly or entirely unconscious fantasies, and to differentiate them from everyday language usage of 'fantasy', meaning conscious daydreams, fictions, and the like. In psychoanalysis, unconscious 'phantasies' are always inferred, but never observed as such. (On 'fantasy' and 'phantasy', see Hook, 1979;Isaacs, 1948, pp. 74-80.) Freud's own use of Phantasie does not, without arbitrary force, lend itself to such a clear-cut distinction. In Phantasie, Freud largely envisaged a process of passage between different psychic systems (from the unconscious via the preconscious to the conscious systems, and vice versa). In this article, only the word 'fantasy' is used henceforth. (On fantasy in Freud's writings, see Laplanche & Pontalis, 1967/1973; on fantasy in the scope of different psychoanalytic approaches, see Levy & Inderbitzin, 2001).

As mentioned above, the term 'fantasy' has a marginal status in the cognitive domain. However, there are synonyms, and in the postbehaviorism era there has been a lot of experimental study on mental imagery, quasi-perceptual experience and daydreaming (see, e.g., Singer, 1966Singer, , 1975Singer, , 1976Thomas, 1999). Brain activity during imagination, for example, has been studied along with evolving methods of brain sciences (Halpern, 2001;Kosslyn, 1994), and-not surprisingly-a daydreaming computer program ('Daydreamer') has been devised (Mueller & Dyer, 1985). Thus, nowadays psychoanalysis and cognitive orientation share an interest in fantasies/imagery, although different kinds (or different aspects) of fantasies/mental images are focused on. Psychoanalysts are interested in fantasies concerning what somebody might do or be, whereas cognitively oriented researchers study matters such as the mental rotation of threedimensional objects. Psychoanalysts, in turn, have used the term 'imago' to refer to an unconscious image or to an early object (or a part-object), including somatic and emotional elements in an object relation, and bodily and instinctual aspects in unconscious fantasy (cf. Isaacs, 1948, p. 93).

In the drawing together of fantasy and the unconscious, the topic of unconscious fantasy appears as a very unlikely meeting point of psychoanalysis and cognitive orientation.

Fantasy: The 'Psychoanalytic' and 'Cognitive' Aspects

As far as the dynamics of the mind is considered, it is easy for a cognitivist to accept the central psychoanalytic idea that our desires are often in conflict with other desires, and with the knowledge we possess. In spite of this, the word 'irrational' does not usually characterize the core fields of cognitive orientation-researchers are mainly interested in competencies, skills and information processing, for example. There is, however, a branch of study that could be said to use fantasies systematically as a research tool. We are referring to philosophers' 'thought-experiments': stories about the twin earth, the blind neurophysiologist Mary and the fictive neurosurgical operations (see, e.g., Chalmers, 1996) that are known by most philosophers. In recent years, cognitive researchers have also taken emotions more and more into account, and along with evolutionary theory, sexuality has been studied a great deal. All this could be seen as steps toward 'hot' and 'wet' aspects of human beings in cognitive research (see, e.g., Dalgleish & Power, 1999.) There are no reasons to doubt that psychoanalysts and cognitivists refer to the same kind of phenomena when using the term (conscious) fantasy/ imagery, respectively: the sequence of events manifests itself in our consciousness as inner speech or image-like (or quasi-perceptual) formations. Cognitive views (Mueller & Dyer, 1985, for example) stress that fantasies (daydreams, imagery) possess the functions of supporting the planning of the future, of learning from successes and failures, and of creativity. They may serve as 'copings' in solving problems and in controlling emotions. Conversely, psychoanalysts (see, e.g., Laplanche & Pontalis, 1967/1973 emphasize that our more or less conscious desires are formulated in a more or less distorted way in fantasies. Thus, it seems that cognitivists and psychoanalysts are interested in different kinds or in different aspects of fantasies.

If we think of a young man's fantasy of building his own house with a fireplace that is almost the same as the one in the house built by his father, we can separate the 'cognitive' and the 'psychoanalytic' aspects of a fantasy.

A cognitivist might say that the fantasy makes previously acquired knowledge accessible to the young man, and in this way it helps him to realize his plans concerning his own house. A psychoanalyst would note that the idea of having his own house and the idea of the man's father seem to be somehow related. He/she would pay particular attention to the fireplace in the fantasy in terms of how and why it would be similar to or different from the one existing in the house built by his father. Psychoanalysts and psychodynamic therapists, as well as psychoanalyzed people, have noted that studying a fantasy and the associations linked with it often reveals something one has not known previously. In our fireplace example, by studying the fantasy more closely, the young man would probably learn something new about what his father means to him.

After having sketched the role of fantasy, imagery and daydreaming in psychoanalytic and cognitive traditions, we will now consider how these matters are thought to exist.

Unconscious Fantasies: How Do They Exist?

Psychoanalysts often describe unconscious fantasies as having 'highly organized mental contents'. This means that the contents of unconscious fantasies are assumed to be quite similar to the contents of consciousness-'ready-made' unconscious ideas. From the cognitive viewpoint, however, this is a problematic presupposition (O'Brien & Jureidini, 2002). (Explicit) memories exist as neural representations in our brain, and at a given moment most of them do not appear in our consciousness. We might say that this is the only non-problematic example of unconscious ideas: memories exist as non-active neural representations of the brain. Similarly, when the term 'unconscious fantasy' refers to a forgotten fantasy, we do not face problems when determining its existence (cf. Baars, 1996). However, when bearing in mind the constructive nature of remembering, even this way of determining the nature of non-conscious ideas is problematic.

Psychoanalysts have not been able to determine how unconscious fantasies exist (Arlow, 1969, pp. 3-4;Beres, 1962;Inderbitzin & Levy, 1990, pp. 116-117;Lyon, 2003). The arguments presented in favor of their existence cannot be accepted in the context of cognitive research. Slap and Saykin (1994) tell us that unconscious fantasies could be seen as film clips, but it is difficult to see how this might be true. Beres (1962, p. 314) advocates the notion of unconscious fantasy by referring to the fact that patients often react to events by rapidly producing a (conscious) fantasy. The quick emergence of a fantasy would not be possible if it had not existed as 'ready-made' in the unconscious. A cognitivist might react to this by asking: 'Are you suggesting that the brain is not quick? ' Freud (1915' Freud ( /1953d) stated that the unconscious can never be studied directly, and that we can know it only through its consequences, or 'derivatives': nobody has ever seen an unconscious idea, and no research tool reveals unconscious fantasies. Does it make sense to argue that X cannot be studied directly, but that there is compelling indirect evidence of its existence? Here it is reasonable to distinguish the matters that we can be conscious of directly from the ones that we can be conscious of only indirectly. We are directly conscious of our contents of consciousness-both present and past: I am conscious of gazing at the computer now, and I am conscious of having listened to Stevie Ray Vaughan's guitar-playing yesterday (a memory concerning this event can be brought to my consciousness). We can also be directly conscious of what we look like (by looking in a mirror), what we know, and-in some cases-what happens in our body (noting a pumping feeling under the chest, for example).

We cannot be directly conscious of our heart, but we have learned that the pumping feeling under our chest is caused by an organ named the heart. We have also learned what affects its pulse, and we may have even learned to control it to a certain extent. Similarly, we are indirectly conscious of the implicit knowledge we possess: we may be aware of its effects. However, compared with the heart, implicit knowledge is difficult to name and to locate (it is implemented in distributed neural networks). Freud's idea on indirect evidence concerning the existence of unconscious matters resembles this way of being indirectly conscious of bodily matters. However, neuroscience has proved that the heart exists, but no neuroscientist has ever seen an unconscious fantasy.

All in all, unconscious fantasies are a cornerstone of psychoanalysis, but it seems impossible to say why we should believe that they exist. Thus, some psychoanalytic researchers have given up the 'realistic' view (Dilman, 1984, pp. 32-45;Laplanche, 1987Laplanche, /1989Solms, 2003).

Mental Imagery: How Does It Exist?

Interestingly, researchers of (conscious) imagery have faced a similar problem when trying to determine the existence of the object of study: when we have a visual or an auditive image, there are no light-waves or sounds (or changes of air pressure) anywhere, and thus it is difficult to say how images exist. Zenon Pylyshyn (2003), one of the leading researchers in the field, states: 'Yet if there is something special about the format in which we think when we have the experience of "seeing with the 'mind's eye'", nobody has satisfactorily articulated what it is, despite some 300 years of discussion' (p. 113). However, there is a lot of empirical research on mental images, and it is clear that many competencies are based on imagery.

There are two main approaches to explaining these competencies: on the one hand, there is 'picture theory' and, on the other, there is 'description theory'. Both approaches hold that there are neural representations in the brain that are responsible for the cognitive operations occurring during imagination. The approaches differ as to what kind of information is thought to be coded in these representations. To put it briefly, according to picture theory, images are coded, and according to description theory, sentences are coded (see, e.g., Thomas, 1999). Despite the fundamental problems involved, there are evident reasons to talk about imagery and to study it.

Nelson Goodman (1990) states, 'talk of "pictures of the mind" taken literally is nonsense' (p. 363). However, he gives a positive answer to the rhetorical question, 'Can I say that there are no mental images and yet say quite intelligibly and truly that I now have in my mind a picture of Capri?' (p. 360). Goodman argues that we should treat mental images in terms of neurophysiological structures, which give rise to certain competencies: an ability to talk and think about matters even when they cannot be perceived (see also Pylyshyn, 2003). Although the contents of mental images (pictures or voices) cannot be pinpointed, it is a fact that we possess an ability to process non-existent matters in our mind. Indeed, the perception of something, on the one hand, and imagining it, on the other, have been found to share certain brain processess (Halpern, 2001;Kosslyn, 1994).

When trying to account for unconscious fantasies, we seem to face problems on two levels: on the one hand, there are general difficulties in determining how conscious mental entities (like imagery) exist, and, on the other hand there are 'psychoanalytic' difficulties in determining how unconscious mental entities exist. Researchers legitimizing their research interest in both imagery and unconscious fantasies refer to indirect evidence: people's (verbal) behavior. If we are to fully understand the significance of the idea of unconscious fantasy to psychoanalysts, we have to study what kind of (verbal) behavior in the analysands provokes it.

The Logic Behind the Psychoanalytic Idea of Unconscious Fantasy

Let us begin to approach the question presented above by asking how psychoanalysts come to the conclusion that one possesses an unconscious mental content. Basically, this happens in a similar way to when researchers conclude that a subject possesses an implicit representation: by noticing certain patterns of behavior. In psychoanalysis, and outside laboratories in general, the patterns are more complicated. Analysts' observational data consist of what analysands tell them, their transference reactions, networks of associations, slips of the tongue, resistances (e.g. missing analytic sessions) and analysts' own counter-transference reactions, for example. Analysts constantly formulate hypotheses about the hidden contents of their analysands. These hypotheses contain ideas such as that an analysand may be (unconsciously) angry with or jealous of a certain person, or that he/she (unconsciously) desires or fears something, and some of them are later confirmed by the analysands, some are reformulated, and some are discarded.

The confirmation of an analyst's hypothesis is a tricky issue, however. Let us suppose that, at first, an analysand thinks that she has almost only warm feelings toward her mother. She has much criticized her analyst and her boss (both female), even when there was no apparent reason. When the analysand talks about traumatic experiences in her childhood, she often blames her father for them (although, for the analyst, her mother seems to be responsible for them as well). On the basis of the pattern that emerges, the analyst makes the following hypothesis: the analysand possesses repressed aggressive ideas about her mother. Later she abandons the idea of having mainly warm feelings toward her mother-instead, she feels disappointed with her mother, and is often angry with her. She may even say that she has previously possessed unconscious hatred toward her mother.

Is the analysand right, and how does she know that? Talk about one's own unconscious ideas makes some sense only when it concerns one's past self. If we think in terms of having an unconscious idea now, either the idea in question is conscious, or we are just guessing (or formulating a hypothesis). However, we have to ask whether we really have privileged access to our past self: how can we know or feel that we have previously possessed an unconscious idea? The answer seems to be: by noticing our past patterns of behavior (see Talvitie, 2003;Wegner, 2002, pp. 171-184). Thus, only patterns of behavior (and some affects attached to them) have been perceived, not an unconscious idea.

Was the analysand unconsciously angry with her mother or not?

We think the answer is both 'no' and 'yes-in a certain sense'. On the levels of 'access consciousness' and 'phenomenal consciousness' (see, e.g., Block, 1995;Carruthers, 2000;Perner & Dienes, 2003), or in terms of explicit knowledge systems (Perner & Dienes, 1999), the analysand was not angry. However, unconscious structures and processes of her brain produced (patterns of) behavior that referred to anger, or that might be interpreted as such.

Where, then, do fantasies fit in? From the point of view of the existence of unconscious matters, there is no crucial difference between an unconscious idea and an unconscious fantasy. Fantasies are just more complicated than single ideas or contents such as 'A will do P, and I will get B,' or 'The occurrence of A leads to the realization of B.' Hypotheses about unconscious fantasies aim to explain more complicated patterns than those about single unconscious ideas or contents.

Unconscious Fantasy and 'Figurative Language'

In the course of the above considerations, a picture of the tension between (clinical psychoanalytic) practice and theory has emerged. On the one hand, the idea of unconscious fantasy has been proved to be a good 'tool' for psychoanalysts. On the other hand, it has also been proved that is is difficult to tell how these fantasies might exist. In his essay 'Do Unconscious Phantasies Really Exist?', Mark Solms (2003), who might be called the founder of neuropsychoanalysis, gives his proposal for relieving the tension.

Solms mainly focuses on introspection and the nature of the existence of conscious fantasies. He holds that such things can be 'seen' only from the (subjective) first-person point of view. It is only in the last four pages that he turns to unconscious fantasies.

He relates (Melanie Klein's) views on unconscious fantasies to Freud's metapsychology (ideas concerning psychic energy, cathexis and psychic tension, for example). For Solms, the term 'unconscious phantasy' and (metapsychological) mechanical explanations are just different points of view about the same thing: talking in terms of unconscious fantasies renders it possible to bring theories concerning the mechanisms of the mind/brain into the domain of the language of experiences (Solms 2003, pp. 102-103). According to Solms, the 'language of unconscious phantasy is just another figurative language' (p. 103).

Freud's ideas about our unconscious life are often praised as his most important contribution. However, we could also grant that status to his notions concerning surprising interrelations between different matters-such as personal history, transference and fantasies in the example presented above. Freud was able to identify that kind of complex dynamics due to the method of study he developed, and he came up with his ideas about the unconscious in order to explain those notions.

Only a little over a hundred years after the publication of Freud's neurophysiological study Project for a Scientific Psychology (Freud 1895(Freud /1953a are we able to present neurophysiological explanations for his notions of complex dynamics. Solms relates unconscious fantasies to Freud's 'mechanical' models of the mind/brain. This idea is elaborated on below: unconscious fantasies are related to 'mechanical' models of presentday cognitive neuroscience. We will begin by considering the kinds of neural mechanisms that give rise to verbal behavior that leads analysts to think of their analysands in terms of unconscious fantasies.

Cognitive Science and the Unconscious

From the perspective of cognitive science, it is difficult to argue in favor of ready-made unconscious ideas, but in any case, there are plenty of uncon-scious matters that affect our behavior: the existence of implicit memory (see, e.g., Graf & Masson, 1993), implicit learning (Berry, 1997) and implicit knowledge (Underwood, 1996) has been shown in numerous research paradigms.

Cognitive science was long criticized for restricting itself to 'cool' cognition and leaving emotional and motivational factors aside. However, in recent years neuroscientists such as Joseph LeDoux (1998) and Antonio Damasio (1999) have published well-known studies in which they show that emotion and cognition cannot be separated from each other. Along with the growing interest in emotions and motivation there is substantial evidence of how emotional and motivational processes take place outside a person's awareness (see, e.g., Chartrand & Bargh, 2002;LeDoux, 1998).

Usually, implicit memory, or, more generally, the effect of non-conscious matters on our behavior and conscious states, is studied in the following way. In the first phase, subjects are presented with stimuli, or they perform a task. The second phase shows that they were not aware of (or do not remember) the stimuli presented, or that they do not know a certain dimension of the task. The third phase of the experiment purports to show that, in the first phase, the subjects acquired unconscious knowledge that later affected their behavior. Non-conscious processing has been shown to shorten reaction times, to affect how we interpret stimuli, to activate goals, and to give rise to emotions (see the references above).

Empirical studies as well as clinical notions concerning people with brain damage (see, e.g., Shimamura, 1993) have shown that we possess implicit, non-conscious representations. The crucial difference between implicit and explicit representations is that the former affect our behavior without giving rise to conscious states. They exist only as neural structures which become activated in certain situations. They do not have content, and are only dispositions of behavior. In the present context, it is also important to note that it is not thought that implicit representations can be made explicit-the unconscious cannot be made conscious (see Talvitie & Ihanus, 2002).

It is not only unconscious representations, but also unconscious neural algorithms that have provoked discussion. Edelman and Tonnoni (2000, pp. 176-190) support us in our approach to 'Freudian' matters such as unconscious intentions, obsessions and repression, in terms of such information-processing algorithms.

The heart of the problem with psychoanalytic ideas concerning the unconscious is the (often tacit) supposition that there is a resemblance between conscious contents and unconscious matters-it is considered possible to bring repressed contents to consciousness. However, conscious ideas (beliefs, fears, desires, fantasies) are wholly different entities from those possessed by the unconscious brain: there are just neural networks in the brain (see Dennett, 1987, pp. 117-202). Thus, we should approach differently the psychoanalytic idea of becoming conscious of the repressed.

Although the repressed cannot be made conscious, we can acquire knowledge about the functioning of different kinds of neural systems of the brain-just as we can acquire knowledge about the functioning of our heart. This idea is put forward below.

Integrating Cognitive and Psychoanalytic Views

There is a clear difference between the psychoanalytic and the cognitive views of the unconscious. Among the non-psychoanalytic researchers, certain neuroscientists (such as Edelman & Tonnoni, 2000) note the relevance of psychoanalytic insights. According to LeDoux (2002):

That explicit and implicit aspects of the self exist is not a particularly novel idea. It is closely related to Freud's partition of the mind into conscious, preconscious (accessible but not currently accessed), and unconscious (inaccessible) levels. However, Freud's terms carry much theoretical luggage that I want to leave behind. (p. 28) Within LeDoux's conceptual framework, 'an unconscious fear' could be seen as follows: without conscious recognition the brain detects a pattern that has previously appeared in a fearful situation. This detection affects the processing of the brain and leads to reactions that refer to fear (bodily reactions, affects, certain kinds of associations and fantasies, for example).

Perhaps Freud's (1911Freud's ( /1953c idea about the two principles of mental functioning is part of the 'theoretical luggage' that does not fascinate cognitive researchers. In any case, on the one hand, we find that the unconscious processing of stimuli supports realistic reactions toward the outer world ('reality principle'), and is thus in the service of adaptation. LeDoux focused on this aspect of unconscious processing in his study of the unconscious detection of signs of danger, for example. On the other hand, the unconscious detection of the characteristics of, say, another person may direct one toward the past rather than the realities of a current state of things. Let us consider the example that the brain detects that the manner of speaking of a perceived person is similar to that of another person met previously. This kind of unconscious detection gives rise to certain feelings, goals, associations and memories. This may be realistic, but the (detected) part does not necessarily match with the whole-although the two people have similar manners of speaking, they may be different in all other respects. In that case, unconscious processing serves as the grounds for unrealistic or, say, imaginary mental functioning.

There is plenty of talk about unconscious meaning in psychoanalysis, although in philosophical debate this talk is problematic. However, it appears sound in the context of the empirical study of unconscious processes and structures, as introduced above: when the brain detects, without the help of conscious processing, a property of a stimulus, and that detection makes a difference to the behavior of a person, it is natural to talk about unconscious meaning.

Cognitive research on the unconscious, as referred to above, has been intensively elaborated in psychoanalytic circles in recent years (see, e.g., Kandel, 1999;Levin, 2003;Solms & Turnbull, 2002;Stern, 1998;Talvitie & Ihanus, 2002).

In elaborating Solm's idea in the spirit of neuropsychoanalysis, we should maintain that 'figurative' talk in terms of unconscious fantasies refers to such unconscious detections, representations and neural routines. The causes of an analysand's behavior can be stated only in terms of neurophysiology (it is the brain that possesses causal power). However, talk about the unconscious structures and processes of the brain makes no sense in psychotherapy. Thus, these neurophysiological matters have to be treated in phenomenal language-when thinking in terms of unconscious fantasies, the analyst transforms neurophysiological matters into a form that makes sense from the point of view of an analysand's experiences. Phenomenal language means so-called 'folk-psychological' terminology: talk about (both conscious and unconscious) desires, beliefs, fears and fantasies. We might say that talk about unconscious fantasies abstracts the functioning of the brain. The third-person 'objective' point of view is enough for most neuroscientists and empirical researchers, but in the talking cure, the perspective has to be that of the first person (cf. Solms & Turnbull, 2002, pp. 4-6).

It seems that, if we are to make sense of unconscious fantasies, concentrating solely on the (ontological) questions 'Do they exist?' and 'How do they exist?' is inappropriate-we must also ask whether the concept is useful. In general, in using a psychological term, one is not (always) assumed to be able to pinpoint the reference to it (from the brain). According to this line of thinking, talk about unconscious fantasies is-in principle-as appropriate as talk about other theoretical entities such as 'schemata', 'internal models', 'scripts' and 'cognitive maps'.

In order to make the theoretical viewpoints presented above more accessible, let us study a fictive clinical example.

Unconscious Fantasy and Psychoanalytic Practice

Let us suppose that, given an analysand's certain patterns of behavior, an analyst has formulated a hypothesis that the analysand possesses the unconscious fantasy that 'My analyst does not respect me, and she may overlook me because of the needs of other analysands and other people.' One day the analyst gets into a car accident, and so the analysand does not find her analyst when she comes for her analytic session. A fantasy comes to her mind: another patient may be in trouble, and the analyst is seeing him in a psychiatric hospital.

Cognitivists (e.g. Currie, 1995) often state the function of fantasies (or imagery) to be simulation: with the help of fantasies, events can be anticipated and one's own activities planned. Is it possible to set the analysand's fantasy into this framework? The fantasy might be true, but it is difficult to take it as the most probable reason for the fact that the analyst is not in her consulting room. Thus, when dealing with this kind of fantasypresented very often in psychoanalytic literature-cognitive viewpoints are difficult to apply: a fantasy seems to be more closely related to one's fears, desires and life history than to possible future events.

The fantasy is studied in the next analytic hour, and the analysand talks a lot about her mother and younger sisters: as the oldest child, she had to take care of herself, and her mother concentrated on her younger sisters. She also found a more general notion (fear) in her mind: like her mother, the analyst would probably not pay special attention to her either.

Thus, we can make sense of the fantasy from the point of view of the analysand's life history. In the specific situation, the analyst's general hypothesis was verified, although the unconscious fantasy she hypothesized cannot be pinpointed anywhere. Thus, we could not say either that an unconscious fantasy became conscious, or that unconscious fantasies (which cannot become conscious, contrary to many analysts' [e.g. Arlow, 1969] view) shape the contents of our consciousness, dreams, metaphors and neurotic symptoms.

Conversely, we should acknowledge that the unconscious processes of the brain affect how we interpret situations (why did the analyst miss the analytic session?), what kind of guesses or suppositions we make (the analyst is meeting another analysand in the hospital), and what kind of conscious fantasies we are disposed to create. Following this kind of thinking, we could say that the analysand possessed preconditions to create a certain kind(s) of fantasy(-ies). 'Unconscious fantasy' should be understood as an analyst's hypothesis about what kind of conscious fantasies an analysand is predisposed to produce.

On Realism and Instrumentalism Concerning Unconscious Fantasies

Along with the considerations made above, problems to do with unconscious fantasy have been given a new form-the ontological question (do unconscious fantasies exist?) has become the practical one: is it plausible or useful to talk about unconscious fantasies? This view is indeed the one Ludwig Wittgenstein held toward psychoanalysts' talk about unconscious matters (see Bouveresse, 1991Bouveresse, /1995. Talk about schemata as well as about both conscious and unconscious fantasies (or images and daydreams, which are terms that are used in the domain of cognitive orientation) is motivated not by their existence, but by practical considerations.

Regardless of how we answer the ontological question, we should ask the practical question whether it makes sense to talk about the brain in terms of unconscious fantasies. The above discussion on the reasons why psychoanalysts say 'yes' here provides an instrumentalist view of unconscious fantasies. Some of the ideas put forward by Mark Solms gave the starting point, but, confusingly, it is not clear whether he is a realist (i.e. believing that unconscious fantasies exist somehow in the mind/brain) or an instrumentalist.

Considering that Solms (2003) maintains that talk about unconscious fantasies is 'just another figurative language', it is surprising that he states, 'what the patient becomes aware of is not the unconscious phantasy itself. Rather, it is a conscious representation (or retranscription) of the phantasy itself' (p. 104). Does he, after all, hold that in the unconscious mind there lie unconscious fantasies? Probably not. Let us take one more citation: '[Unconsious fantasies are] descriptions of the things that actually do make up the inner workings of our minds, and which really do cause us to think and feel our conscious thoughts and feelings' (p. 104). Is he suggesting here that it is possible-at least in principle-to pinpoint in the brain certain structures and processes that are responsible for each unconscious fantasy?

Solms's view cannot be made sense of without noting his leaning on Freud's original metaphor concerning consciousness (see Solms, 1997). Most neurophysiologists and philosophers currently maintain that the brain causes conscious states, but Freud approached consciousness in terms of perception: just as we may perceive things around us, we may also perceive 'mental things' of our minds. Some 'unseen' mental things may be brought into the scope of our inner perception (the relieving of repression), while some remain forever outside it. Solms seems to hold that unconscious fantasies deal with the latter case. Given Solms's conception of consciousness, his view of unconscious fantasies appears somewhat difficult to grasp: mainly he seems to be an 'instrumentalist' (talk about unconscious fantasies is just figurative language that is useful for psychoanalysts), but occasionally he emits a 'realistic' tone (unconscious fantasies exist as dispositions of the brain, but they can never be perceived).

In the scope of the current article, the latter approach appears problematic, for two reasons. First, choosing to study consciousness through the metaphor of perception creates problems for the project of neuropsychoanalysis: cognitive views become (slightly) difficult to apply. Second, given the fact that the brain is a very complex system functioning in a non-linear way (see Kelso, 1995), it is difficult to poinpoint from it the disposition to produce a certain fantasy. Thus, we would advocate a plainly instrumentalist view of unconscious fantasies, and suggest that Daniel Dennett's works (e.g. 1987) may serve as a foundation for such a view (see Talvitie, 2003).

Discussion

In psychoanalytic circles, there have always been debates about whether psychoanalysis falls in the domain of science or hermeneutics. Currently, psychoanalysis has a higher status in the latter domain-in the departments of cultural studies and film studies, for example, Freudian (or neo-Freudian) insights are often leaned on, whereas currently it is difficult to find a psychoanalytically oriented professor of psychiatry.

Neuropsychoanalysis might be seen as a revival of 'scientific' psychoanalysis. In terms of the lines of thought put forward above, science and hermeneutics do not appear as alternative frameworks for psychoanalysisinstead, it seems that one cannot give up either of them. From the 'scientific' side, it is clear that in many cases psychoanalytic ideas are tied to the matters studied by cognitive neuroscience: both neuroscience and empirical research on implicit knowledge have approached the unconscious processes of the brain, and reached results that are highly relevant for psychoanalytic ideas concerning the unconscious.

Psychoanalysis is composed of theories that can be evaluated-at least partly, and among others-from the point of view of cognitive neuroscience. Hermeneutics or, say, handicraftsmanship becomes important when we study psychoanalysis as a practice aimed at helping individuals to overcome their psychic problems. Although the psychoanalytic situation (an analysand and an analyst on different sides of the couch) could be seen as a certain kind of laboratory, it is very different from situations of empirical research: experiments cannot be repeated, variables cannot be manipulated, for example. The clinical practice of psychoanalysis and other psychotherapies cannot rest solely on scientific knowledge either: not every 'intervention' of a therapist can be based on empirical research, for many reasons.

In terms of cognitive neuroscience, in psychotherapy one aims at altering the functioning of the brain by speaking (LeDoux, 1998, pp. 263-265). Psychotherapists in general hold that it makes no sense to speak about the brain in therapy, and psychoanalysts in particular have found it useful to talk about matters such as fantasies, and to think in terms of metaphorical relations between distinct matters. These kinds of insights set (at least) the other foot of psychoanalysis onto the domain of hermeneutics. Above, a link between scientific (neurophysiology) and hermeneutic (thinking in terms of unconscious fantasies) aspects of psychoanalysis was created-we suggested that talk about unconscious fantasies should be seen as a clinically useful way to talk about the brain.