Patricia Garfield (1999a) has sketched out a new 12-part classificatory system for what she claims are universal dreams. "I propose that there are l2 basic dreams, each of which have positive as well as negative versions" (Ibid., p. 1). The system includes such categories as chase or attack, vehicle trouble, being lost or trapped, and being menaced by the dead. The 12 general categories include varying numbers of subcategories, so there are actually 29 categories in all. In constructing her system, Garfield drew on several sources, including dreams she has collected in various parts of the world, her own 50-year dream journal, and the literature on dream content. In addition, she also relied upon a system devised by folklorists for classifying folktales; she thinks this system "can be adjusted to classify dreams because of the strong resemblance between dreams and folktales" (Ibid.).
Garfield believes her classification system is necessary because other systems have not "taken firm and satisfactory international hold" (Ibid.). She notes that the Hall/Van de Castle system comes "nearest" to being "a generally accepted system," and says it is "helpful for research," but she finds it lacking because it "leaves the individual dreamer with the question, 'what does my dream mean?'" (Ibid.).
As a strong advocate of the Hall/Van de Castle system, I am very pleased that Garfield would single it out as widely accepted and useful for research. But is she right that there is a need for a new system to supplement it, such as the one she suggests? And could her system actually tell individual dreamers what specific dreams mean? This paper suggests that the answer to both of these questions is "No." It first of all shows that the Hall/Van de Castle coding system is far more comprehensive than Garfield's system and readily encompasses all of it except for four low-frequency categories of minor importance. The paper secondly discusses the many methodological problems inherent in her system, none of which is present in the Hall/Van de Castle system. Thirdly, this paper uses findings from studies of dream content to express doubts concerning the adaptability of a system for classifying folklore to the study of dreams. Finally, it argues that no system of classification could make it possible to tell a person what a specific dream means; to the degree that Garfield offers any interpretations of any of her universal dreams, it is on the basis of the unstated theory that dreams express our self-conceptions, and that they often do so through the use of metaphor.
Garfield's Categories for Universal Dreams
The first analysis in this paper concerns the degree of overlap between Garfield's 29 categories and the Hall/Van de Castle system. The short answer is that almost all of her system fits within just 5 of the 10 general categories in the Hall/Van de Castle system. In order of importance, her categories are encompassed by the Hall/Van de Castle categories for Misfortunes and Good Fortunes, Social Interactions, Successes and Failures, Characters, and Emotions. Rather than go through all of Garfield's categories in sequential order and great detail, let me make some general comments and provide a few examples before turning to the methodological problems within her system.
Most of Garfield's categories have to do with what are called misfortunes and good fortunes in the Hall/Van de Castle system. Misfortunes and good fortunes both concern events that just happen to people "out of the blue" for no seeming reason. Unlike aggressive or friendly acts, they are not due to the actions of the dreamer or other dream characters. Unlike successes or failures, they are not due to striving by the dreamer or other dream characters. It is a Misfortune, for example, for a character to become ill or lost, or to lose a possession. It is a good fortune to suddenly become well, to find money, or to be in a bountiful environment. Misfortunes, then, are bad things that happen to people without intention, and good fortunes are lucky, unexpected, wonderful things that happen to people.
The following 10 of Garfield's 29 categories are misfortunes in the Hall/Van de Castle system: injury or death; vehicle trouble; house or property damage or loss; losing valuable possessions; being in a house on fire; falling or drowning; being naked or inappropriately dressed in public; having a machine or telephone malfunction; natural disasters; and being lost or trapped. Five of her categories qualify as Good Fortunes in the Hall/Van de Castle system: meeting a star; healing or rebirth; flying; being in a place of great natural beauty: and discovering "marvelous" new spaces, such as a whole new room in a hoiuse or treasures in the attic. (Depending on the nature of the interaction, the meeting with the "star" also might be coded as a friendly interaction.)
Dream elements that are coded as aggression, friendliness, or failure in the Hall/Van de Castle system account for most of the few remaining dimensions of Garfield's system. Her categories for chase/attack and being menaced by the dead are aggressions. Her categories for being embraced/loved and guided by the dead are friendly interactions. Her category for doing poorly on a test or other performance is a failure. In addition, a few of her categories have overlaps with Hall/Van de Castle categories for characters and emotions. Animals, prominent people ("stars"). and dead people are characters who have specific designations in the Hall/Van de Castle system. Her categories for flying "joyfully" and traveling "happily" overlap with the happiness dimension of the five-part schema for classifying emotions.
There are only four dimensions in the Garfield system that are not readily classified into one of the major categories in the Hall/Van de Castle system. They are "driving skillfully," "making a home or property improvement," "being dressed handsomely," and "operating a machine or telephone easily." For these dream elements, more than one category of the Hall/Van de Castle system would have to be used. Such a necessity has the virtue of making it possible to establish empirically how frequently the separate elements actually appear together, but the fact remains that none of these rare elements fits easily into the Hall/Van de Castle system.
"Driving skillfully" would be a location change (L) within the overall activities category; if the vehicle being driven is mentioned, it would be coded under travel (TR) in the objects category. As for "home or property improvement," it has no obvious categorization without more details on what is actually happening in the dream; if the dreamer is working on the house, it would be a physical activity (P), and the house would be coded under architecture/residential (AR) in the objects category. "Being dressed handsomely" would be classified under clothing (CL) in the objects category and as a positive evaluation (E+) in the modifiers scale. "Operating a machine or telephone easily" would be coded as a physical activity (P) within the activities scale and possibly as low intensity (I-) on the modifiers scale; a machine would be coded under tools (IT) in the objects category, a telephone under communications (CM) in the objects category.
Given the great overlap of the two systems, and Garfield's very different goals, are there reasons to favor the Hall/Van de Castle system over Garfield's? Generally speaking, there are two. First, her system has many methodological problems. Second, the Hall/Van de Castle system has coding features and established findings that are a good basis for determining what, if anything, is universal in dream life.
The Methodological Problems in Garfield's System
There are five general methodological problems inherent in Garfield's system. All of them would have to be resolved in a satisfactory fashion before intercoder reliability or the validity of the system could be determined. First, it is a daunting task to try to develop a system to classify dreams as a whole because may dreams easily fit in more than one category. It is perfectly plausible that a dream would have a chase, an embrace, and damage to a possession, for example, thereby fitting into three of Garfield's 29 categories. There are two possible ways to deal with this problem. One is to develop rules for deciding which aspect of a dream is primary in order to place it in one of the categories. The trouble with this strategy is that it throws away information by removing all trace of what are defined as the secondary and tertiary aspects of the dream. For this reason, a single category for each dream has limited usefulness. The other possibility is to classify each dream in as many categories as are necessary to encompass all of its content, which means the system might end up looking much like the Hall/Van de Castle system for classifying elements within dreams. Either way, much work on decision rules needs to be done.
In her follow-up commentary, Garfield (1999b, p. 9) suggests that her 12 basic themes could be seen as "elements" that combine to produce "compounds," and that dreams could be Classified under more than one category. Using the example. of a friend's dream in which he/she "was driving my car with no brakes down crowded city streets, while naked, when I got lost...", she says that such a dream could be identified as "car trouble + naked in public + lost," or in numbers as "3.1 + 7.0 + 11.0." Such multiple classifications move in the direction of traditional content categories and tend toward the "minute detail" she hopes to avoid, but with 'little possibility for rigorous quantitative development.
Second, as things now stand, the system only encompasses a small part of all dream content. Since the Hall/Van de Castle system is a descriptive one that was purposely expanded to the point where it has a category for virtually everything that is found in dreams, it can be used as a benchmark on this point. As shown earlier, Garfield's system essentially deals with only three of Hall and Van de Castle's 10 general categories-- Misfortunes and Good Fortunes, Successes and Failures, and Social Interactions. It has little or nothing to say about settings, objects, food and eating, elements from the past, or the most frequent types of characters in dreams. Emotions appear only as part of one or two other categories, such as "traveling happily" or "flying joyfully." Thus, it seems that the Garfield system might miss many of the cross-cultural similarities in dreams that have been found with the Hall/Van de Castle system (Domhoff, 1996). This suggests that an expansion of the system is needed, which means once again that the system might come to resemble the Hall/Van de Castle system it is intended to supplement.
Third, the Garfield system is based on dimensions that are defined in terms of polarities, such as "injury or death vs. healing or rebirth," or "being naked or inappropriately dressed in public vs. dressing harmoniously and handsomely." However, not all the polarities that Garfield poses are obvious, or free of unstated theoretical assumptions about what constitutes a dimension. For example, it is not clear to me that falling and drowning are always the opposite of flying. Many of the falling dreams we have collected involve falling off a building or through a hole in a floor or walkway, not falling from the sky or out of nowhere. Moreover, not all flying dreams are positive in emotional tone. To take another example, being trapped or lost does seem like the obvious opposite of finding a new room in your house.
Even with polarities that seem plausible, such as one defined by aggression and friendliness, it is not a foregone conclusion that either aggression or friendliness is best understood through placement on such a dimension. Argument and evidence would have to be presented to show why this makes sense instead of dealing with each social interaction separately, as the Hall/Van de Castle system does. In the Hall/Van de Castle system, there are first of all ways to analyze aggression and friendliness separately, and then there are also ways to determine the dreamer's mix of friendly and aggressive interactions with each dream character and with dream characters in general. Before the Garfield dimensions are adopted, they should be compared with the Hall/Van de Castle approach to social interactions to see which yields the most useful results. Such a comparison could be done very efficiently by applying her dimensions to the normative male and female dream reports used by Hall and Van de Castle. These dreams, along with the codings for each dream, are available to all dream researchers at http://www.dreambank.net/ (Schneider & Domhoff, 1999).
Fourth, several of Garfield's categories are too general and imprecise in that they mix elements that should be kept separate and then aggregated later for some analytical purposes. For example, she has several categories that include both aggressions and misfortunes, two that include both characters and social interactions, and two that include both emotions and activities, and one (the "embrace or love" category) that seems to encompass both friendly and sexual interactions. Consider, for example, how she defines her category for "injury or death:" "You or another dream character (often a loved person) is injured, killed, or dies. Usually the cause is accidental. The villain or threat is not emphasized; injury or death just happens" (Ibid., p. 25). But as numerous Hall/Van de Castle findings show, it is useful to keep aggressions and accidents(misfortunes) in separate categories. Being killed is one thing, suffering an accident is another. Or take her category for "house and property damage or loss," which includes dreams in which "your pocketbook or wallet is missing, misplaced, or stolen" (Ibid., p. 25). Losing your pocketbook is one thing, having it stolen is another. Finally, she has a category that includes natural and man-made disasters that range from earthquakes to atomic war. I think an earthquake is one thing (misfortune), and an atomic war is another (an aggression).
Turning to another way in which Garfield's categories are too general and imprecise, her categories called "menaced by the dead" and "guided by the dead" are really subcategories of her chase/attack and embrace/love dimensions; they should not be separate categories just because the attacker or befriender is dead (or recognized as someone who has died, if that is what is meant by this category). As stressed earlier, it is better to have separate categories for distinct elements such as social interactions and characters, and then to determine empirically how often aggression or help comes from specific types of characters.
As a final example of this point, two other categories, "flying joyfully" and "traveling happily," should be considered as a second step after flying, traveling, and emotions have been considered separately. Done in that way, flying "fearfully" or traveling "apprehensively" also can be studied, along with the degree to which flying and traveling are linked with positive emotions. Making these kinds of distinctions becomes even more compelling when it is realized how easy it is to combine categories as a second step. For example, after looking at aggressions and accidents separately, or types of travel and emotions separately, it is then possible to look at them together. Put another way, it is much easier to aggregate Hall/Van de Castle categories than it would be to disaggregate Garfield's categories. This is the basic reason why she is wrong to build a system on general motifs rather than on what she calls "the easily measurable, the minute detail" (Garfield, 199b, p. 21). Such critical turning points in scale development are overlooked when people new to content analysis start out to create a new classification system.
Fifth, the Garfield system does not have a way to take into account sequences of events, such as when an aggression is followed by a friendly embrace, or sickness is followed by healing. By comparison, the Hall/Van de Castle system has a detailed set of rules for coding any "consequences" that follow after Social Interactions, Misfortunes and Good Fortunes, or Successes and Failures, making it potentially more valuable in any attempt to classify dreams into types. The various consequences are rare in dreams in general, and we therefore seldom code for them, but they may be more frequent in the kinds of dreams that interest Garfield. For example, if a dream character is sick or dead, and then experiences healing or rebirth, then this is a Misfortune followed by a Good Fortune. A character who becomes lost and then is guided to safety by a dead person would have suffered a Misfortune that is followed by a Friendly Interaction. In other words, an emphasis on the sequencing or interaction of categories is the best way to study some of the thematic aspects of dreams, and would have to be added to the Garfield system to make it more useful.
As this section shows, there are many methodological problems that would have to be dealt with before the Garfield system could be tested for reliability and validity. Solving these problems is no easy task, as Hall and Van de Castle (1966) learned to their chagrin in two years of revising and discarding various of their subcategories. For this reason alone, it might make more sense to try to accomplish Garfield's classificatory goals with an established system. But there are also several positive reasons for sticking with the Hall/Van de Castle system rather than trying to supplement it with an entirely new and untested system.
Hall/Van de Castle Findings
In addition to the several methodological virtues of the Hall/Van de Castle system pointed out in the previous section, there is one other reason for building on the Hall/Van de Castle system. It already has produced a large number of findings on dreams that can provide a basis for studying what, if anything, is universal in dream life, and for judging the likelihood that various of Garfield's categories are as universal as she claims. These findings begin with the normative frequencies for some of the dream elements considered by Garfield to be universal. For example, our computer search of the codings for the 500 male and 500 female dreams used by Hall and Van de Castle (1966) to establish their norms revealed that 11.6% of the men's dreams and 8.8% of the women's dreams had a chase (A6), attack (A7), or murder (A8) in which the dreamer was a victim. This finding supports Garfield's claim that the chase/attack dimension may be a universal one. On the other hand, we found no instances of dead characters menacing or helping any other characters. In fact, there were only 7 dead characters in the 1000 dreams: 6 in dreams from women, 1 in a dream from a man. (There are no norms for the appearance of deceased people who are alive in the dreams.)
In addition to the normative findings, which can be searched for any possible contingency between two or more dream elements, there are numerous cross-cultural findings with the Hall/Van de Castle system that can be brought to bear on the question of universal dream elements. These findings not only come from industrialized democracies such as the United States, Switzerland, and Japan, but from a few dozen small traditional societies in many different parts of the world that were studied by anthropologists between 1920 and 1980. Moreover, there is much more that could be done with Hall's large collection of cross-cultural dreams if money were available to take the dream reports and his codings of them out of musty file cabinets and enter both the reports and the codings into our DreamBank. Not only could the codings be searched for the frequencies of such elements as chase (A6) and attack (A7), but the dream reports themselves could be searched for specific words and phrases using the fast and efficient search engine developed especially for dream research by Adam Schneider (Schneider & Domhoff, 1999).
For now, however, we have to settle for codings of these dreams by Hall for many categories, as summarized in Domhoff (1996), and by Van de Castle (1983, 1994) for animals in dreams. Hall's findings show, for example, that chase and attack dreams are as ubiquitous as Garfield assumes they are. In addition, they show that there are a great many dreams with the kinds of misfortunes denoted by several of Garfield's categories: being lost (M1), falling (M2), losing valuable possessions (M4), and suffering injuries (M5).
On the other hand, the rarity of certain findings cast doubt on the universality of a few of Garfield's categories. The best example here concerns her belief that having a "magical" animal "friend" is worthy of a category in her system. Such dreams supposedly involve "mythological" or "talking" animals, or "encountering enchanted beasts, such as a precious tame, talking bird, a snake who wears a golden crown, or a beautiful woman with a fawn's body" (Garfield, 1999a, p. 25). However, Van de Castle's (1983, 1994) work suggests that such animals must be very rare in dreams; the majority of interactions with animals in the dreams of people of all ages anywhere in the world are highly negative. On this point, it is also noteworthy that there are more aggressive than friendly interactions with animals in the dreams of both men and women in the Hall/Van de Castle normative sample, and only one instance of an animal befriending a dreamer; the specific findings are presented in the following table.
The examples presented in this section are only a small portion of what could be done with Hall/Van de Castle findings in determining the universality of a wide range of dream elements and dream themes, including the 29 categories that make up Garfield's system. For now, the meager findings for some categories raise questions about Garfield's decision to draw upon a system for classifying folklore in creating her categories. This issue is addressed in the next section.
Dreams and Poetic License
As noted in the introduction, Garfield says that she drew in part upon a system for classifying folklore in creating her system for studying universal dreams. Judging by the rarity of such themes as friendly mythical animals and dead characters menacing dreamers in the dreams I have access to, I think she has relied too heavily on folklore, literature, and myth in developing her system. This inference is strengthened by the fact that early in her article she mentions a dream by Gilgamesh from 4,000 years ago (Garfield, 1999a, p. 1).
Although myths and folktales do include themes of mythical animals, and shamans often talk about dreams in which they are aided or impeded by dead people or spirits, neither folktales nor claims by shamans are evidence that people actually have such dreams. One of the great problems in studying dreams is that there is no way to check up on specific claims by individuals about what they allegedly dream. The fact that the word "dream" can be used for visions, hopes, and waking fantasies as well as our nightly dreams only adds to the confusion about what is a dream and what is not. The mystique surrounding dreams is so great that the phrase I dreamt"" can be used as a platform for any claim a person may wish to make, however implausible. This means that assertions about dream content in myth and literature, or in discussions with shamans and visionaries, cannot be taken at face value. There is really only one way to decide the matter, which is to collect and study thousands of dreams from living individuals of all ages in as many countries as possible. Within that context, it can be suspected that atypical dreams reported in any type of literature or shamanic lore are as likely to be the product of poetic license as they are to be actual dreams.
The possibility that poetic license and folklore have contributed to Garfield's system raises serious doubts about its relevance for studying dreams. These doubts can only be resolved through careful empirical studies. In addition to more studies of dreams from sleep laboratories, classrooms, and dream journals, I think there may be another way to approach this problem, which is to ask people of all ages to make up a dream that will seem real to the people who read it or hear it. Such a study could bring out the stereotypes that people have about dreams, especially when such dreams are compared with a "real" dream from the same participants or with the Hall/Van de Castle normative findings.
There are also studies that could be done to see if the dreams of interest to Garfield are frequent enough to be of any great importance in the overall study of dreams. Such studies would compare what people report on questionnaires about allegedly universal themes in dreams with what they actually dream about. To the degree we have such comparisons, they show wide disparities. For example, in a search for 10 allegedly common dream themes in a sample of 983 dream reports from two-week dream diaries from 126 students at the University of California, Santa Cruz, in 1992, we found almost no dreams containing one of the 10 themes, even though many of the same students reported on an earlier questionnaire that they remembered having such dreams at least once. To take the case of flying dreams, 59% of the men and 54% of the women said they recalled having such a dream, but there were only 5 flying dreams in the dream diaries, or 0.5% of the total dreams reported. There were only two dreams of being lost, two of taking an examination, and one of falling (Domhoff, 1996, p. 198). I recognize that frequency is not the same as universality, but at the least such findings suggest that the study of common, allegedly universal, dreams might not be very productive.
Clearly, much more empirical work would need to be done to decide just how typical elements like friendly mythical animals and menacing dead people are in dreams. My point for now is that findings with the Hall/Van de Castle system provide the best starting point, calling into question Garfield's adaptation of a system developed for studying the very different, not similar, material found in folklore. Until such time as systematic empirical studies show otherwise, findings such as those provided by Van de Castle (1983, 1994) on the role of animals in dreams should set the standard.
Can Garfield's System Interpret Specific Dreams?
One of the advantages Garfield claims for her system is that it can help people understand what a specific dream means. "The Universal Dreams" concept," she writes, "may bridge research and popular needs" (Garfield, 1999a, p. l). I doubt that any classificatory system can achieve this goal, and I see no evidence that it is accomplished in Garfield's system. To the degree that Garfield offers any hints in this direction, they are asides that amount to metaphoric glosses that may or may not be appropriate for a specific individual. For example, in relation to dreams of "driving skillfully," she categorically states that "few dreams fall into this category, but when they do they are extremely significant, suggesting easier access to skills for coping with difficult life situations" (Ibid., p. 25). In the case of dreams of "dressing harmoniously and handsomely," she claims "these dreams sometimes refer to satisfaction with the appearance of our bodies, or may refer to a situation in which we feel that we 'fit' well." But which is it for the dreamer? She provides no guidance as to how we would know whether to give the "satisfaction with body" or the "fit the situation well" interpretation to an individual dreamer (Ibid., p. 26).
In fact, Garfield's interpretations have nothing to do with classifications and categories. Instead, they are based on two theoretical assumptions that are not made explicit or demonstrated. The first and most basic is that dreams express our self-conceptions. In the case of the "driving skillfully" dream, the self-conception expressed is one of having "skills for coping with difficult life situations" (Ibid., p. 25). In the case of the "dressing harmoniously and handsomely" dream, the self-conception is also a positive one, either thinking of our body positively or thinking of ourselves as fitting in well socially. In making this point, I am not being critical of the theory. In fact, it is a very plausible one first put forward, coincidentally enough, by Calvin Hall, 46 years ago, when he sketched out a cognitive theory of dreams in which dreams are hypothesized to express conceptions of oneself, significant others, and the social environment (Hall, 1953b). Instead, I am emphasizing that the issue is a theory, not a classification system, but I also would add that the evidence for the theory implicitly espoused by Garfield is not so convincing that everyone accepts it.
The second theoretical assumption underlying Garfield's interpretations is that self conceptions are often expressed metaphorically. Once again, this is a very reasonable theory, and it is again once one put forward long ago in another article by Hall (1953b). This theory is also the real basis for Jungian, Gestalt, and phenomenological dream interpretation, even though Gestalt and phenomenological theorists vigorously assert that they do not accept the idea of symbolism in dreams. It is also the stock-in-trade of most popular dream theorists and dream educators, even though they claim they do not have a theory, but simply rely on intuition and common sense. Metaphoric interpretations now have wide currency, but there is not much, if any, systematic evidence for the very interesting idea that dreams are at least sometimes metaphoric in nature.
Even if we assume for the moment that both of these theoretical propositions turn out to be true, there is still the problem of deciding which of several possible metaphoric interpretations is correct for an individual, such as a person who finds himself or herself "dressing harmoniously and handsomely" in a dream. Here Garfield and many others might say that the decision should be based on other information about the dreamer, but I think outside information about the dreamer creates more problems than it solves. This is partly because the "interpretation" is no longer based on the dream itself, but also because both the dreamer and the interpreter are very susceptible to self-delusion in such a situation because our minds search for closure and we are overconfident of our powers of judgement. From my point of view, the only way someone can provide evidence that they really can interpret a dream is to do a "blind analysis," that is, to make the interpretation without knowing anything about the dreamer.
Where, then, can information for making the decision come from? I think there are at least two possibilities. First, it could come from other information within the given dream that is being interpreted. For example, studies could be done of large samples of dreams containing the theme of "dressing harmoniously and handsomely" in order to see if there are "contingent" elements that distinguish between dreams that should be interpreted in terms of the person's feelings about her/his body and dreams that should be interpreted in terms of fitting into a social situation. Second, the necessary information could come from other dreams from the same person. In this case, a "representative sample" of the person's dream life would be used to provide a context for narrowing down the meaning of the specific dream. Based on our empirical studies of subsamples from long dream journals, "representative sample"would be defined as at least 100 dream reports that reflect everything the person recalled during the time period the dreams were written down.
In making these comments about the possibilities for dream interpretation, I do not want to leave the impression that I think we are at the point where we can interpret an individual dream. We can extract information about a person from a series of dreams, but until we can be confident that dreams really do express our conceptions, and sometimes use metaphoric thinking in so doing, we cannot tell a person what his or her dreams mean.
Garfield's proposed classificatory system for universal dreams has no advantages and numerous methodological disadvantages in comparison to the Hall/Van de Castle system, and this is without even considering the large amount of time and effort that goes into refining categories so that an acceptable level of reliability can be attained. In addition, the Hall/Van de Castle system can encompass all but a few minor dimensions of Garfield's system in a more precise fashion and provide a normative basis for future studies of such "themes" as being lost (M1), being naked in public (M4), failing an exam (FL), being attacked (A7), and flying (GF).
There are, as many studies have shown, some universal dream elements, but it is very doubtful that a classificatory system for universal dreams could take us beyond this well-established finding. Moreover, some of Garfield's allegedly universal dreams, even if they are universal, are so rare as to be of minor importance in developing a theory of dream meaning. Driving skillfully, doing home improvements, and dressing handsomely are not frequent enough to be featured in a system of universal dreams. Still other allegedly universal dreams, such as those of friendly or enchanting mythical animals, must be taken with a grain of salt until such time as large samples of representative dreams have been studied.
When it comes to understanding the meaning of dreams, neither Garfield's system nor the Hall/Van de Castle system can substitute for a sound theory, but the Hall/Van de Castle system can be of great use in developing such a theory.
Go back to the Dream Library index.