Why did empirical dream researchers reject Freud? A critique of historical claims by Mark Solms

G. William Domhoff

University of California

Mark Solms (1997) made a major contribution to dream research through his neuropsychological study of 361 clinical patients suffering from a wide variety of brain injuries. By asking these patients if they had noticed any changes in their dreaming since their injuries and then comparing their answers with brain scans and psychological tests, he was able to develop hypotheses concerning which parts of the brain are and are not necessary for dreaming. He then buttressed one of his key findings by resurrecting a large literature from the 1940s and 1950s on lobotomized schizophrenics, which shows that most of them lost their capacity to dream. It is likely they did so because the operation was done in one of the areas that Solms found to lead to a cessation of dreaming in some of his patients -- the white matter beneath the ventromesial surfaces of the frontal lobe (also known as the frontal limbic white matter).

The previously uncharted "neural network for dreaming" suggested by Solms's findings receives significant support from several positron-emission tomography (PET) scan studies of rapid eye movement (REM) and non-REM (NREM) sleep in normal adults, which show a pattern of reactivation and continued deactivation during REM that paralleled his findings with the clinico-anatomical method (Braun et al., 1997; Braun et al., 1998; Maquet et al., 2000; Maquet et al., 1996; Nofzinger, Mintun, Wiseman, Kupfer, & Moore, 1997). The combination of Solms's lesion studies and brain imaging studies has led to a renewed interest in empirical dream research and revisions in activation-synthesis theory, the first neuropsychological theory of dreaming (Hobson, Pace-Schott, & Stickgold, 2000b; Hobson, Stickgold, & Pace-Schott, 1998). Thanks in part to Solms's work, then, it is now possible to think in terms of a neural network for dreaming that can be connected with findings from developmental studies in the sleep laboratory (Foulkes, 1982, 1999) and with systematic studies of dream content (Domhoff, 1996, 2003).

Unfortunately, however, Solms has tried to use his empirical work as evidence for the viability of Freudian dream theory without taking into consideration any of the many systematic empirical studies inside and outside the laboratory setting that call into question every hypothesis about dreams put forth by Freud (see Domhoff, 2003, pp. 136-143, for a full discussion of this evidence). In the process, he has reinterpreted the history of laboratory dream research, and especially the history of the controversy over NREM dreaming, in a very Freudian-serving way. He has also wrongly charged that many past theorists failed to take seriously subjective reports from awakenings in the sleep laboratory, but then ignores widely known reports from laboratory awakenings with children ages 3-8 that contradict every aspect of Freudian theory (Foulkes, 1982, 1999).

The purpose of this article is twofold. First, I discuss the history of the controversy over REM and NREM dreaming that Solms misunderstands. This recounting shows that Freudian dream theory could not have been rejected by most dream researchers for the reasons Solms describes. Second, I present a brief summary of the main empirical findings that explain why Freudian theory is not considered viable by most dream researchers.

Solms's version of history

Solms presents his version of the history of laboratory dream research in one chapter in a book that attempts to integrate the findings in many areas of neuropsychology with psychoanalytic theory. Its goal is to create a new field called neuropsychoanalysis (Solms & Turnbull, 2002). The book is coauthored with Oliver Turnbull, a neuropsychologist, who is also the secretary of the new Society for Neuro-Psychoanalysis founded by Solms and his coworkers. Because Turnbull's work focuses on visuospatial disabilities, laterality, and waking disorders, Solms had the major responsibility for the chapter on dreams (M. Solms, personal communication, August 20, 2003).

According to Solms and Turnbull (2002, p. 186), the discovery of REM sleep wrongly led to the rejection of Freudian dream theory because "the early investigators concluded that REM sleep (read: dreaming) is an entirely 'mindless' activity." These early dream researchers therefore readily accepted the claims of activation-synthesis theory, which asserts that dreams are caused by random signals arising from the tegmental region of the pons. According to the original version of this theory, the forebrain merely reacts to these signals, so it is simply "making the best of a bad job in producing even partially coherent dream imagery from the relatively noisy signals sent up to it from the brainstem" (Hobson & McCarley, 1977, p. 1347). Thus, dreams are motivationally neutral and without a purpose, and therefore do not have the importance that Freudians attach to them.

This theory, Solms and Turnbull (2002) continued, led to a debate about the viability of Freudian dream theory at a panel during the meetings of the American Psychiatric Association in 1976. Although the association was "still dominated by members sympathetic to psychoanalysis," a vote of the members present at the debate went "overwhelmingly against Freud -- suggesting that this was the end of the road scientifically speaking for Freud's account of the mechanisms of dreams" (Solms & Turnbull, 2002, p. 190). It is therefore "no exaggeration to suggest that the tide turned decisively against psychoanalysis in America at that fateful 1976 meeting" (Solms & Turnbull, 2002, p. 191).

The equation of REM sleep and dreaming supposedly was accepted by most dream researchers despite the fact that Foulkes (1962) had shown that a significant number of NREM awakenings lead to dream reports that are similar to the reports from REM awakenings. Solms and Turnbull (2002, p. 195) then said that Foulkes's work was "neglected" because too many dream researchers do not take "subjective reports" seriously. This state of affairs prevailed until 1997, when Solms published the clinico-anatomical findings he developed between 1985 and 1990. By showing that there were patients with intact pontine regions who did not dream, and some patients with pontine lesions who continued to dream, Solms claimed he demonstrated the importance of the forebrain in generating dreams. His work thereby led to a reconsideration of the work by Foulkes and others on NREM dreaming. More importantly from Solms's point of view, his work reestablishes the credibility of Freud's theory of dreams. If dreaming had not been dissociated from the pons, it "might still be possible to claim that dreams are 'mindless' and 'motivationally neutral' " (Solms & Turnbull, 2002, p. 196). However, because of his findings, along with work on the dopaminergic system by other investigators, what is known about dreaming is now "broadly consistent with Freud's psychoanalytic theory of dreams" (Solms & Turnbull, 2002, p. 215).

As shown in the next section, this story is highly inaccurate. Empirical dream researchers never thought that dreaming in normal adults is a mindless activity, even at the time (the late 1950s and early 1960s) when most of them believed that all or nearly all vivid dreaming occurred in REM sleep. In fact, many of them were motivated in their work by an interest in Freudian dream theory, including the most important pioneer in dream research, William C. Dement, who did his postdoctoral work on dreams in New York with the psychoanalyst Charles Fisher. Dement later wrote that he went to medical school with the intention of becoming a psychoanalyst because he was an "ardent disciple" who had "read and re-read The Interpretation of Dreams" (Dement, 1992, p. 25; Foulkes, 1996, p. 610).

Moreover, dream researchers took the work on NREM dreams by Foulkes and many others very seriously. They had largely abandoned the strict correlation of REM and dreaming by the last half of the 1960s, well before the arrival of activation-synthesis theory. Thus, few if any dream researchers were persuaded by activation-synthesis theory when it was propounded in the mid-1970s, so this theory could not have had anything to do with the widespread rejection of Freudian dream theory by that time. Far from rescuing Foulkes's NREM findings, Solms's account leaves out any consideration of the many laboratory findings by Foulkes and other empirical dream researchers. The cognitive theory of dreams favored by Foulkes (1985) and others (Antrobus, 1978, 1991; Cavallero & Foulkes, 1993; Hall, 1953a, 1953b) is completely ignored.

The history of the REM/NREM controversy

Contrary to Solms, there was immediate discussion of Foulkes's findings when they appeared in 1962, especially because they supported findings on the significant number of NREM dream reports in an earlier paper by a team of researchers at the Downstate Medical Center in Brooklyn (Goodenough, Shapiro, Holden, & Steinschriber, 1959). There also was an extensive study by Kamiya (1961) showing the same results. For all intents and purposes, the controversy was settled on the basis of two studies in the laboratory setting created by Allan Rechtschaffen, one of the founders of the new field of empirical dream research (Foulkes & Rechtschaffen, 1964; Rechtschaffen, Verdone, & Wheaton, 1963). As Foulkes (1996, p. 612) later recalled in an overview of the new dream research from 1953 to 1993, the combination of Rechtschaffen's stature in the field and the findings from his lab "ultimately, if not easily, carried the day." To the degree that there was resistance to these findings, it had several roots. For one thing, the equation of REM and dreaming created financial support from government funding sources that hoped the new discovery might substantiate the long-held belief among psychiatrists and philosophers, suggested well before Freud's time, that there was a link between dreaming and psychosis. There also was the hope that the apparent REM basis for dreaming would be a window into the mind-body problem (D. Foulkes, personal communication, September 24, 2003). As far as the substantive resistance to abandoning the strict correlation of REM and dreaming, it tended to come from the more Freudian-oriented researchers, which is the opposite of what Solms's historical account might lead readers to expect. This reluctance was related to three factors that at the time seemed crucial to future studies of dreams and to a full testing of Freudian dream theory.

First, there was a strong belief that the eye movements during REM were closely related to dream content, which meant that it might be possible to have a behavioral indicator that linked directly to dream content. Although the main published study on eye movements and dream content was based primarily on anecdotal evidence with small samples (Roffwarg, Dement, Muzio, & Fisher, 1962), the findings had standing because they came from the laboratory of the leader in the field, Dement, and because they were supported by changes in other physiological variables that seemed to relate to changes in dream content (see Dement, 1965, pp. 173-184, for a detailed summary of the work by Roffwarg, Dement, and others on the relationship between REM and dream content).

Second, there was also the fact of "REM rebound" after participants experienced what was then called dream deprivation, which seemed to support the idea of a need to dream in just the way Freudian dream theory might expect (Dement, 1960; Dement & Fisher, 1963). Third, there was preliminary evidence that the rise and fall of penile erections during REM related to dream content. According to C. Fisher (1966, pp. 541-542), at the time an active participant in dream research as well as in the psychoanalytic community, he correctly predicted seven out of eight instances of detumescence on the basis of aggressive elements in the dream content, such as the dreamer being bitten by a dog, snake, or lion. In addition, he correctly predicted five out of six instances of "a sharp, sudden increase in tumescence," most of which involved erotic dream content. Here, then, was the possibility of studying the sexual symbolism that is so ubiquitous in dreams according to Freudian theory.

However, none of these high hopes withstood the test of further studies. The claims about eye movement patterns could not be replicated in a very thorough study (Moskowitz & Berger, 1969). Then the eye movements were found to be quite stereotypic and usually different from waking eye movements (Jacobs, Feldman, & Bender, 1970, 1972), a conclusion later supported in an even more detailed study (Aserinsky, Lynch, Mack, Tzankoff, & Hurn, 1985). Nor could the dream deprivation studies be replicated. As Vogel (1975, p. 749) concluded after summarizing the many studies, "REM sleep deprivation is not dream deprivation and is not harmful to schizophrenics, depressives, or healthy subjects." Finally, no further claims were made that penile erections are related to dream content; in fact, it was found that they sometimes occur in NREM sleep before REM periods begin (Hursch, Karacan, & Williams, 1972).

During the middle and late 1960s, then, two processes were occurring simultaneously -- the breakdown of the relationship between physiological variables and dream content within REM periods, and the increasing likelihood that dreaming also occurred outside REM sleep. The combination proved irrefutable even for those who had tried to maintain the strict correlation between REM periods and dreaming. Thus, by the early 1970s at the latest, there were few laboratory dream researchers who believed that dreaming was confined to REM sleep.

The final nail in the coffin, if one were needed, came from a carefully conceived study undertaken with the "extensive help" of a social psychologist, Robert Rosenthal, who was an expert on creating and assessing "expectancy bias" (Herman, Ellman, & Roffwarg, 1978, p. 92). The experimental design created by Herman et al. (1978) included two laboratory assistants who served as the "experimenters" and four groups, each with 10 participants. The key variable was the false belief that some participants were being given a drug that had been found to increase dream recall from 20% to 92% in NREM sleep and from 86% to 99% in REM sleep.

For the first group, both the experimenters and the participants thought that the recallenhancing drug had been administered. For the second group, the experimenters thought they were administering the drug, but the participants were told that the "pills" were to reduce skin resistance to improve EEG recordings. For the third group, the experimenters thought they were administering a mere placebo, but the participants were told they were receiving the recall-enhancing drug. For the fourth group, the experimenters thought they were administering the placebo, and the participants thought the pill was to reduce skin resistance.

The study showed that raising the expectancy of dream recall in either the experimenters or the participants could increase the frequency of dream reports from NREM awakenings, but the findings with the fourth group revealed a large amount of NREM dreaming that could not be accounted for by expectancy biases. The investigators concluded that their results were "in line with those investigators who found substantial amounts of NREM mentation and a relatively small REM-NREM discriminability" (Herman et al., 1978, p. 90). Significantly, they ended their discussion with a comment that is the opposite of what might be expected based on Solms's account: "The hope that one stage of sleep, or a given physiological marker, will serve as the sole magic key for vivid dream mentation has all but faded from view" (Herman et al., 1978, p. 92). From that point forward, the argument among those who collected dream reports in the sleep laboratory primarily concerned whether or not there should be a correction for the length of reports in determining if there were some or no differences between REM and NREM reports (Antrobus, 1983; Domhoff, 2003; Foulkes & Schmidt, 1983; Hobson et al., 2000b; Hunt, 2000).

Based on this history, it can be seen that there were virtually no empirical dream researchers who were predisposed to accepting Hobson and McCarley's (1977) article that is focused on to the exclusion of all else by Solms. Instead, it was natural scientists and psychiatrists outside the dream research community who found the ideas in the article attractive, primarily because of the growing disenchantment with psychoanalysis due to its relative failure in helping seriously disturbed patients, especially in comparison with the success of the antipsychotic and antidepressant medications that were introduced in the late 1950s and 1960s. The ideas also appealed to philosophers of consciousness, such as Churchland (1988) and Flanagan (2000), who had widespread influence. Then, too, the ideas were widely presented as the counter to Freudian dream theory in introductory psychology textbooks (Squier & Domhoff, 1998), and they were picked up by the press, because they represented an all-out challenge to Freudian theory.

However, despite the widespread attention to activation-synthesis theory, it had far less impact within the dream research community. In fact, the new theory came as a great surprise to most researchers, many of whom were put off by the fact that the Hobson and McCarley article did not refer to a single study based on empirical dream research in the human sleep laboratory. In suggesting that everyone accepted the new theory, Solms does not seem to understand that it was as much of a challenge to laboratory dream researchers as it was to the psychoanalysts.

Vogel (1978) in effect answered for laboratory dream researchers with a very detailed refutation based on three different types of evidence. A psychiatrist and dream researcher who gradually abandoned a Freudian view in the face of the empirical findings, Vogel thanked Rechtschaffen for helping him in writing his rejoinder. First, Vogel showed that attempts in the late 1960s and early 1970s to relate the "phasic" events of REM sleep -- such as bursts of eye movements, muscle activity, or bursts of theta waves -- to unusual dream features through immediate awakenings revealed no strong correlations; this conclusion was reinforced in later studies (Antrobus, 2000; Pivik, 1986, 2000). Second, he pointed out that there is evidence that the forebrain plays a role in regulating REM sleep, a conclusion that has received renewed support in recent years (Morrison, Sanford, & Ross, 1999; Salin- Pascual, Gerashchenko, & Shiromani, 2000). Third, Vogel noted the solid evidence for dreaming outside REM sleep, with a particular focus on the dreamlike mentation at sleep onset before there had been any possibility of a REM period (Foulkes & Vogel, 1965; Vogel, 1991; Vogel, Barrowclough, & Giesler, 1972). It is noteworthy, and a hopeful sign, that activation-synthesis theorists have adapted their theory to incorporate these and other new findings (Hobson, Pace-Schott, & Stickgold, 2000a; Hobson et al., 2000b; Hobson et al., 1998).

Vogel (1978, p. 1534) concluded his critique by suggesting there are probably multiple sources for the level of cortical activation that makes it possible for dreaming to occur. Thus, the idea that there may be two or more dream instigators is not unique to Solms (1997, 2000), although he has marshaled further evidence to support the hypothesis.

As this account shows, Solms could not possibly be correct in his explanation as to why Freudian theory came to be ignored. To understand the rejection of the Freudian theory of dreams, it is necessary to briefly overview the empirical evidence that has accumulated against it. None of this evidence has been cited by Solms in any of his publications, in which Freudian dream theory is discussed in a positive light.

The evidence against Freudian dream theory

The essence of Freudian dream theory is that dreams are a disguised attempt at wish fulfillment that serve the important function of preserving sleep in the face of episodic pressure from motivational urges like sex and hunger (Freud, 1900). They are disguised by a process called the dream-work, which operates at the behest of what Solms calls the censorship. Because the dream-work makes most adult dreams unintelligible, it is necessary to have patients free associate to each aspect of a dream to understand it. As the rest of this section discusses in summary form, there is strong evidence that contradicts all of these claims.

The Function of Dreams

When it comes to the function of dreams, laboratory dream research calls Freud's claims into question in a number of ways. First, dreams turn out to be longer, more frequent, and more regular than Freud (1900, p. 377) imagined when he said they were like a "firework that has been hours in the preparation, and then blazes up in a moment." These findings contradict the hypothesis that dreams are brief reactions to episodic urges. Second, Foulkes developed unexpected evidence, first in a longitudinal study (Foulkes, 1982), then in a cross-sectional replication (Foulkes, Hollifield, Sullivan, Bradley, & Terry, 1990), that children do not dream very often until they are age 8-9, yet they sleep soundly. In both of these studies, based on many hundreds of awakenings, median reporting from REM periods is only 20%, and the reports are very brief. Moreover, 2 boys with very poor visuospatial skills had similar low reporting rates at both ages 11-13 and 13-15, so it is not easy to claim the children are simply forgetting or do not have the language skills to report dreams (Foulkes, 1982, pp. 180-181 and 225-226).

Third, Freud's functional theory is refuted by Solms's own finding that some patients do not dream but nonetheless are able to sleep. Solms and Turnbull (2002) said that the questionnaire Solms administered to patients showed that those who lost dreaming were more likely to report "disrupted sleep" than those who had not lost dreaming. They wrote that his study "supports Freud's sleep-protection theory at the required levels of statistical significance" (Solms & Turnbull, 2002, p. 214). However, the actual numbers show that only 49% of the nondreaming patients reported disrupted sleep, as compared with 32% of those who retained dreaming. There are far too many nondreamers without sleep problems to salvage a theory that strongly asserts that dreams preserve sleep.

Solms and Turnbull (2002) recognized that this issue must be settled in the sleep laboratory, and they wrote that the "critical test" will be carried out "soon" (p. 213). But one such study already has been done, and it is cited by Solms in his 1997 book. In this study, a team of Canadian researchers studied dream recall from REM sleep over 2 nights in 13 lobotomized patients (Jus, Jus, Villenueve, et al., 1973). Only 10% of the awakenings on the first night led to any indication of "dream presence," a category that includes instances in which the person claimed he had been dreaming but could not recall the content, as well as actual dream recall. The percentage was even lower for the second night (6%; Jus, Jus, Villenueve, et al., 1973, p. 290). Overall, 9 of the 13 participants reported no dreams on either night. Two participants said on one or more awakenings that they had dreamt but could not recall the content, and 2 participants reported some content on one or more awakenings. These figures are far lower than those for the control group of nonlobotomized schizophrenic patients. The research team then repeated the study with 10 lobotomized patients who were given reserpine (Jus, Jus, Gautier, et al., 1973). The researchers again found low rates of recall similar to those in the first study and concluded that reserpine does not enhance dreaming or dream recall.

For my purposes here, the important finding from these studies is that there is no evidence that the nondreaming patients could not sleep or that they slept less well than the control group. In fact, the researchers explicitly asked all participants in the morning about how they had slept: "There is no notable difference in the reports concerning sleep between the two groups of patients. All lobotomized and non-lobotomized patients reported in the morning that they had slept" (Jus, Jus, Villenueve, et al., 1973, p. 282). In addition, the fact that participants experienced REM sleep that could not be distinguished from that of the control group (Jus, Jus, Villenueve, et al., 1973, p. 288) and had to be awakened from it is a refutation of Freud's functional theory. Solms cannot invoke these patients as evidence for nondreaming and still claim that dreams are necessary to preserve sleep.

Because the lobotomized patients had damage to the dopaminergic system that Solms believes is the trigger to dreaming, he might want to argue that dreaming is unnecessary when this system is not operational. However, that would still leave him with the problem that his patients with parietal lobe injuries did not report dreaming either. Thus, even more clearly in this case, because the motivational system is intact, he cannot claim that these patients do not dream and at the same time assert that dreaming is necessary to preserve sleep.

Moreover, there is evidence that some normal adults do not dream. In a study of 12 adult men who said they never dreamed, several of them reported no dreams after 4 to 13 awakenings from REM periods over a period of 3 nights (the exact number of nondreamers is not reported -- only that the range was from 0% to 100% in laboratory recall; Butler & Watson, 1985, p. 825). The nondreamers tended to score lower on a test for visuospatial skills, the same variable that best predicted low or no dream recall in Foulkes's (1982, 1999) longitudinal and cross-sectional developmental studies. More recently, Pagel (2003) conducted a laboratory study of 16 people who reported on an intake questionnaire at a sleep disorders clinic that they never had experienced a dream. Each person spent one night in the sleep laboratory and was awakened two or more times. In a total of 36 REM awakenings, none of the participants reported a dream. By comparison, 3 of 12 people who said they rarely dream reported dreams on 3 of 32 awakenings. The two groups did not differ on any polysomnographic, demographic, or clinical factors.

Based on the several types of evidence presented here, Freud's sleep-protection theory is discredited in two different ways. On the one hand, dreams are far longer and more regular for most people than predicted by his theory that dreams are brief reactions to episodic urges. On the other hand, there is a minority of people who are able to sleep even though they do not dream, including young children, Solms's patients, lobotomized schizophrenics, and at least a few normal adults. On the basis of this evidence, some cognitive dream researchers agree with the activation-synthesis theorists that there is at present no evidence that dreams have any function at all (Antrobus, 1993; Domhoff, 2003; Foulkes, 1993).

Wish Fulfillment

Freud (1900) claimed that "wish-fulfillment is the meaning of each and every dream" (p. 106). Because wishes are usually disguised in adult dreams, he started with anecdotes from his own children, claiming that the dreams of children are easy to understand and "invaluable as proof" of his theory (Freud, 1900, p. 102). The most famous of his examples concerns his 19-month-old daughter, Anna, who called out in her sleep "Anna Freud, strawberry, wild strawberry, scrambled eggs, mash" after going without eating that day due to an upset stomach (Freud, 1900, p. 104). However, Foulkes's (1982) longitudinal laboratory study found that the few dreams that children under age 5 report are bland in content and static in imagery. There is no evidence of wish fulfillment in any of them. Moreover, laboratory research also shows there are frequent microawakenings during sleep, and that sleep talking often occurs during such awakenings (Arkin, 1981; Boselli, Parrino, Smerieri, & Terzano, 1998; Mathur & Douglas, 1995).

Present-day Freudians still can claim that free associations must be obtained before it is certain that the laboratory dreams collected by Foulkes are not based in wishes, but it is noteworthy that Freud did not think such associations were necessary. In any event, the burden of proof is now clearly on the Freudians. What is clear is that they cannot claim sleep-talking episodes as evidence for their claims. The famous Anna Freud dream overheard by Freud probably was not a dream.

The nightmarish dreams that very often accompany posttraumatic stress disorder (PTSD) also provide a direct challenge to Freud's theory of wish fulfillment (Barrett, 1996; Hartmann, 1984). Freud did not discuss such dreams in The Interpretation of Dreams (1900), but he later conceded the main point in the face of the "war neurosis" dreams suffered by combatants in World War I, which is that "it is impossible to classify as wish fulfillments the dreams we have been discussing which occur in traumatic neurosis, or the dreams during psychoanalysis which bring to memory the psychical traumas of childhood" (Freud, 1920-1922/1955, p. 32). While this concession clearly undermines the theory that all dreams are wish fulfillments, Freud (1933, p. 29) also claimed that the "exception does not overturn the rule." Instead, he argued that PTSD dreams do not contradict his theory because they draw on a deeper level of the mind, one "beyond the pleasure principle" that shapes most dreams. His only modification of his theory was to say that dreams are a disguised "attempt" at wish fulfillment (Freud, 1933, p. 29). He then concluded that "unconscious fixation to a trauma seems to be the foremost among these obstacles to the function of dreaming" (Freud, 1933-1936/1961, p. 29).

However, such a resolution of the issue is not satisfactory. The accumulation of findings on the repetitive dreams of PTSD since Freud wrote suggests they cannot be so easily isolated from dream life in general. More people have them, and they persist longer, than Freud's dismissal implies. They are suffered by victims of natural disasters, traffic accidents, rape, and assaults (Barrett, 1996). They are experienced by 15%-20% of the women and men who served in the Vietnam War, and not only by those veterans who saw direct combat (Hartmann, 1984; Kramer, Schoen, & Kinney, 1987; Van der Kolk, Blitz, Burr, Sherry, & Hartmann, 1984). Far from being an "exception" to Freud's theory, the dreams of PTSD show that wish fulfillment dreams are only a subset of all possible dreams.

As Freud anticipated, anxiety dreams and punishment dreams also have been pointed to as evidence against his wish fulfillment theory. However, he replied that anxiety dreams that awaken the dreamer simply show that the censor has failed to disguise the wishes enough to make them acceptable. In the case of punishment dreams, he said that the wish came from the censoring agency within the personality. "With the division into wishful dreams, anxiety-dreams and punishment dreams, we have kept our theory intact," concluded Freud (1933-1936/1961, p. 27). Because these ideas are plausible within the context of his theory, anxiety and punishment dreams are not convincing evidence against it. However, as already shown, no theory-saving claims can be made in the case of the unexpected frequency and persistence of PTSD dreams. Nor can the wish fulfillment theory be readily adapted to the bland and static dreams of early childhood. Thus, these two types of dreams are the best refutation of the wish fulfillment theory as a general theory.

Do Dreams Disguise or Express?

As noted earlier, Freud claims the wishes in adult dreams are difficult to uncover because they have been disguised by the dream-work. However, thousands of awakenings in the laboratory show that dream reports are usually understandable. As early as 1968, one team of researchers summarized their findings from 635 REM awakenings of 58 normal adults as follows:

Laboratory dreams generally involve clear, coherent accounts of realistic situations in which the self is involved in mundane activities and preoccupations. 90% would have been considered credible descriptions of everyday experience....Both the prosaic nature of content and the rich phenomenological similarity to waking experience found are in keep- ing with the results of early introspectionist studies, but discrepant from the usually accepted conception of dreaming. (Snyder, Karacan, Tharp, & Scott, 1968, p. 375) (For a detailed discussion of the results from this study, see Snyder, 1970.)

A later study from Rechtschaffen's laboratory came to the same conclusion (Dorus, Dorus, & Rechtschaffen, 1971). As Foulkes (1985) summarized based on a large body of findings, including several of his own studies, dreams are a reasonable simulation of the real world, and they usually concern everyday issues. They are novel constructions, but they usually do not involve highly improbable events. This continuity with waking conceptions and concerns extends to the emotions in dreams, which contradicts one of Freud's key claims about the dream-work. According to Freud (1900, p. 299), the emotions in dreams are often inappropriate to the content because the contents have been transformed by the dream-work while leaving the emotions "unaltered." However, a detailed laboratory study suggests that the emotions in dreams are "overwhelmingly appropriate to the dream content" (Foulkes, 1999, p. 68; Foulkes, Sullivan, Kerr, & Brown, 1988).

In addition, despite early claims to the contrary by Domhoff and Kamiya (1964), dreams collected in the laboratory do not differ very much in content from those written down at home by the same participants when the proper controls are introduced (Foulkes, 1979; Heynick & deJong, 1985; Strauch & Meier, 1996; Weisz & Foulkes, 1970; Zepelin, 1972). Moreover, a reanalysis of the extensive home-laboratory comparison by Hall (1966), based on 272 REM dream reports and 120 home reports from 8 young adult males, added to the evidence for the similarity between home and laboratory dream reports. It shows that effect sizes were small even when there were statistically significant differences, except in the case of the larger percentage of physical aggressions in home dream reports (Domhoff & Schneider, 1999).

The similarity between laboratory and home dream reports lends weight to a number of findings revealing that dreams collected outside the laboratory setting usually are far more continuous with waking fantasy and personality test results than might be expected from Freudian claims about the effects of the dream-work. These studies are summarized in two detailed assessments of all the evidence for and against Freudian dream theory by S. Fisher and Greenberg (1977, 1996). In addition, quantitative studies using the rigorous and reliable Hall and Van de Castle (1966) system for the study of dream content have shown regularities relating to age, gender, and culture, and that a person's waking conceptions and concerns can be predicted with a sufficient number of their dream reports (Domhoff, 1996, 2003; Hall, 1947). Taken together, these findings suggest that dreams reflect or express more than they disguise. This is also the conclusion of a careful psychoanalytic study in which dreams were collected from 2 participants in the laboratory and then compared with material from their therapeutic sessions and structured interviews (Greenberg, Katz, Schwartz, & Pearlman, 1992).

Contrary to the Freudian claim that the dream-work makes it difficult to obtain much information about dream meaning without the aid of free associations, two different studies suggest that free associations do not add anything to what independent raters can infer about a person based on several of his or her dreams. In the first of the two studies, 12 male and 12 female college students wrote down 10 of their dreams during the course of several weeks, then free associated to each element in each dream in a comfortable room that simulated a therapeutic setting. Using a rating scale for a range of personality categories relating to Freudian theory, the rater made approximately the same inferences with or without the free associations (Reis, 1959). The second and more comprehensive study used dreams, free associations, and transcriptions from psychotherapy sessions with 1 male and 2 female pa- tients to identify the patients' main personal conflicts. The raters' judgments based on the dreams alone were not changed by providing them with free associations (Popp, Luborsky, & Crits-Christoph, 1992).

Finally, it needs to be noted that S. Fisher and Greenberg (1977, 1996) were unable to find any convincing nonclinical studies of any of the processes that are said to make up the dream-work. If there are such processes, they are not likely to be "peculiar to the life of dreams" and "far more remote from the model of waking thought than even the most determined belittlers of the psyche's feats of dream-formation have thought," as Freud (1900, pp. 328-329) claimed. Instead, they are more likely to be the same types of figurative thinking that produce metaphor, metonymy, conceptual blending, and irony in waking life (Lakoff, 1997, p. 90), although there is as yet little or no good evidence to support this interesting possibility (Domhoff, 2003).

In short, a very large body of literature contradicts the claim that dreams are difficult to decipher, and thereby calls the idea of the dream-work or censorship into question. It shows that much dream content is coherent, understandable, and readily related to waking concerns. The main "bizarreness" in dreams is sudden scene changes (Sutton, Rittenhouse, Pace-Schott, Stickgold, & Hobson, 1994), which is not usually thought of as a product of the dream-work. These findings still leave a significant amount of dream content to explain, some of which may be due to figurative thought, but much of which may be meaningless filler (Domhoff, 2003, pp. 128-134; Foulkes, 1985, 1999) or the product of the unusual "structural" properties of the dream state (Hobson et al., 2000b). Whatever the exact amount of remaining dream content that is or is not meaningless, however, the important point for now is that the burden of proof is on those who claim hidden meaning to demonstrate their hypothesis with systematic empirical evidence.

To his credit, Solms makes no attempt to defend the concept of censorship on the basis of neuropsychological work. In fact, in a discussion of the relative inactivity of the prefrontal cortex during REM sleep, he noted that Freud "may have been wrong" about censorship because "the apparent illogicality and bizarreness of dreams may be due to the inherently 'regressive' nature of the dream process," which means there may be "no need to introduce the additional function of censorship" (Solms & Turnbull, 2002, p. 215). He concluded that "For now, we shall have to rely on purely psychological techniques to assesses the validity of this aspect of Freud's dream theory" (Solms & Turnbull, 2002, p. 215). However, Solms did not refer to any of the studies in the large psychological literature on dreams that have been used in this article to question the idea that dreams have meanings that are hidden from ordinary waking consciousness by the censorship and the dream-work. This literature shows that Freud's ideas about the dream-work are as unviable as his claims about the role of dreams in preserving sleep.

Conclusion

As this critique shows, Solms has misunderstood the history of dream research in a very Freudian-serving way. As also argued, Solms has failed to deal with the many different kinds of studies that refute Freudian dream theory on every point where it has proved to be testable, not all of which have been presented in this article (see Domhoff, 2003, pp. 136-147, for refutations of other claims, such as the role of "day residue" and the origins of significant speech acts). By keeping the focus narrowly on the Freudian and activation-synthesis theories as the only viable alternatives, while ignoring the findings that have emerged over the past 40 years from human sleep laboratories, Solms misses an opportunity to contribute to the search for a new and better theory. At the least, he should widen the dialogue so it includes cognitive theories and the findings from content analysis studies. Beyond the general idea that the brain becomes "activated" or "aroused" during dreaming, an idea that is now accepted by all theorists, there is nothing else to salvage from Freudian dream theory. It is time to abandon Freud's theory and use Solms's clinico-anatomical findings as one building block in the development of a new theory of dreams.

References

Go back