DREAMS AND AGING
*Marco Zanasi,M.D.

This work evaluates the influence age has on dreams reports.
The rather limited data to be found in publications would seem to demonstrate that the aging process leads to a reduced ability to remember and recount dreams (although the incidence is reduced as far as women are concerned, Giambra et al., 1996). This datum has been interpreted not so much as a consequence of a weakening of the memory as the sign of a diminished interest in dreams, (Strunz, 1993). As far as dream content is concerned, the data from the publications is divided on the subject but would seem to demonstrate a change in emotional themes, in the sense that elderly subjects' dreams are characterised by a higher incidence of nightmares (Salvio et al. 1992) and, on the other hand, a reduction in emotion (aggression, friendship and activity), Watermann, 1991. Other studies report a greater kindliness in the dream content and an emotional content showing greater contentment (Blick et Howe, 1984). More recent studies suggest that the ability to dream is an acquired cognitive skill that depends in part on the development of the neural network responsible for spatial construction situated in the parietal lobes and that children's clearer memory of dreams is linked to their visuospatial capacities (Foulkes, 1982, 1983,1993, 1996a, 1996b, 1999, Solms, 2000, Domhoff, 2001). On this basis, a study on how the aging process might interact with these neurophysiological processes seemed of great interest.
METHODS
The study covered 148 subjects who were all over the age of 70 (age average 75,87, Standard Deviation 8,42) in good health and 151 subjects between the ages of 18 and 25 (age average 22,45, Standard Deviation 3,23). The elderly subjects were recruited from Old Peoples' Clubs and Universities for Senior Citizens during lectures (the Old Peoples' Clubs and Universities for Senior Citizens are institutions providing recreational and study facilities for the over-65). The control subjects were students from the "Tor Vergata" University, Rome. The research team gave out some general information about the aim of our research and people showing interest in the experiment were contacted a week later. All subjects were asked to recount the last dream they could remember and this was tape-recorded and faithfully transcribed as an accurate working copy. All participants were asked to sign an informed consent form approved by the Ethical Committee of the "Tor Vergata" University, Rome.
In order to analyse the dreams according to the Jungian vision (which looks at the dream as a text produced by the dreamer's unconscious while he/she sleeps; Jung, 1945) we used processing techniques deriving from textual analysis (Gigliozzi, 1997).
If we consider the dream as a text (that is to say, as a well-knit whole or "something woven", to go back to the etymology of the term), then a fortiori we can consider the dream that is reported verbally as a particular form of text which transforms the oneiric experience into an objective product. The dream while it is being dreamed is experience, not text. Our memory of that experience, whether we report is or not, is the text of the dream. So the dream becomes a text the moment the initial experience of it has ended, just as a waking experience can become a text as soon as we are able to reflect on it as "something that happened" to us (Killroe, 2000).
A verbal account of the oneiric datum is the only means we have at our disposal for representing the oneiric experience. We could not produce an account of the dream if the dream itself were not already a textual unity. If, following Danesi (1998), we consider the text a "weaving together of elements taken from a specific code and ordered together so as to communicate something", we are in fact constructing a text when we remember a dream, whether or not we recount it. That the reported dream is a text seems clear from the fact that it has a beginning, a middle and an end. Formally speaking, it has spatio-temporal limits in the sense of a consistent narratological structure (Killroe, ibidem).
The reported dream is also a text that, like a description, will reflect the narrative structure of the oneiric experience. The term "narration" refers, according to Toolan's definition, to "a perceived sequence of events that are linked between themselves in a non-casual manner". The fundamental elements of a narrative have been proposed by Chatman (1978). He divides narrative into two fundamental parts: the "story" and the "discourse". The story is the content or chain of events (i.e. the actions and occurrences) and excludes those elements that could be described as contextual variables (i.e. the characters and setting). The discourse is the form of expression or the means by which the content is communicated. In simple terms, the story represents "what" is described in a narrative and the discourse is concerned with "how" (Chatman, ibid.). In this work we studied the "Discourse" of the dream.
There are numerous studies supporting the belief that reported dreams are a faithful representation of the dream itself. Jung seems to presuppose the adequacy of the dream report as a valid object of textual inquiry, (Jung, ibidem).. Numerous studies that have analysed the contents of dreams show that oneiric content is, in general, analogous to waking thoughts (Kramer et al., 1975). Experiments in which stimuli administered during sleep were shown to have been incorporated into the dreams confirm the existence of a relationship between the oneiric experience and the reported dream. (Kramer et al. 1983). Other experimental research has confirmed that there is a significant similarity between oneiric experience and the reported dream (Roffwarg et al., 1962; Taub et al,1978). More recently Kramer (1993) hypothesizes that verbal accounts accurately represent the original oneiric event.
Proceeding from such a starting point, we used some of the textual analysis techniques developed in relation to the analysis of literary texts for an analysis of our material (Gigliozzi, ibidem). In particular, the oneiric text was evaluated bearing the following aspects in mind:
1. The composition of the text and its character definition
2. The speech's temporal organisation: i.e. the tenses used in the dream report. The narrator of a story or episode can choose between two alternatives: he/she can either state the facts by following the order in which they occurred within a referential (or pseudo-referential) universe or manipulate the narrative's temporal sequences. The latter option has a considerable impact on the organisation of the text insofar as the fabula's sender alters its chronological order as he/she composes his/her speech ("anacronia", in Greek). Indeed, the fabula only exists as an abstraction that cannot be reconstructed empirically, in the sense that it cannot be reduced to a perfect mono-dimensionality whereas speech is, by its very nature, linear.
3. The emotional organisation: the narrative text does not set out a story in an objective and linear fashion but is, in some way, organised by the sender to fit the receiver. The sender programmes the moments and ways in which the data can be received and the story reconstructed by the receiver, as well as the latter's emotional responses. To this end, he/ she can choose how to represent the story and make use of certain artifices: the use of anacronia, particular ways of using quantitative elucidation, "coup de théâtre", narrative paralipsis (passing to one side) and ellipsis (the omission of information) and the attribution of an emotional charge to determinate syntagms. The sum total of such artifices constitutes the speech's emotional aspect or "seiemic"1 narrative level, just as we can define the narrative unities belonging to that level as the "seiem", from the Greek verb seío , meaning "I upset" or "I excite". An analysis of the speech's "seiemic" level can be understood in two different ways: in the broad sense, as the global analysis of the text's formal organisation seen in a "seiemic" perspective or, in the narrower sense, as the analysis of the speech's emotional unities or "seiems".
4. The semantic fields. Semantic fields are groups of words (nouns, verbs, adjectives or adverbs) that are used to describe a particular situation, environment or body of objects constituting a part of our everyday life and creating precise associations in our minds. A semantic field is the area of meaning that a body of words regarding a given subject has.
In particular, the following parameters were examined (Table 1):
* The presence or absence of an observation that defined the place in which the oneiric scene took place and, in the case of such a presence, the further specification of the type of space defined (open or enclosed).
* The definition of the narrative's context or of what could be defined the setting for the oneiric narrative, paying particular attention (in the case of a well-characterised context) to the descriptive or emotional quality of such a definition.
* The presence or absence of chronological observations contributing to a setting of the scene in which the action takes place.
* Linearity or lack of it in the narrative sequence (eg. the presence or absence of flashbacks or of tearing in the narrative texts's continuity and consistency etc).
* The narrative speech's structure or the prevalence of direct or indirect speech or of descriptions given from a position outside the narrative sequence.
* The cast of characters or definition of the dreamer's position as well as that of other possible actors in the oneiric scene.
* Clarification or lack of it regarding the dreamer's emotional state (fear, anger, anguish etc).
* A definition of the situation represented as fantastical or realistic.
* Uniformity or lack of it in the narrative's verb tenses and, where they are uniform, the temporal allocation of the action (present or past).
* The number of words used to compose the narrative was, moreover, counted for every dream.
The frequency distributions for the dream's descriptive variables were analysed and studied both individually and in their correlations. A two-dimensional Contingency table was created for every pair of variables and the absolute frequencies for each line/column intersection were entered.
The two-dimensional distributions were processed using the Chi2 test in order to evaluate the characteristic distributions.
The length of the dreams' texts (both for elderly subjects and control subjects) was compared under the various sub-headings created by the descriptive variables. The comparison test was conducted by analysing the variance from a chosen classification criterion (the ANOVA test).
The T-Test was used to analyse the difference between the number of words in the elderly subjects' dreams and that in the young peoples' dreams.
SEMANTIC FIELDS
We choose to investigate the modification (if it occur) of the various sensory modes in dreams and their modifications with aging.
Relatively little is known about the various sensory experiences in dream reports that occur without known external stimuli (Calkins, 1893; Weed & Hallam, 1896; Hacker, 1911; Kobler, 1913; Bentley, 1915; Knapp, 1956; Snyder, 1970; McCarley & Hobson, 1979; Zadra et al. 1998); as pointed above recent studies suggest that the memory of dreams is linked to visuospatial capacities (Foulkes, 1982, 1983,1993, 1996a, 1996b, 1999, Solms, 2000, Domhoff, 2001).
The semantic fields were evaluated by assessing all the lemmas relating to the senses (sight, hearing, smell and taste) and measuring the frequency of their occurrence in the dreams of the two groups examined.
A study of semantic fields faces the problem of polysemy. The computer is only able to supply a list of the frequency with which the lemmas potentially linked to the semantic field arise in the text. It is not able to appreciate the ambiguity linked to polysemy and the fact that many words have multiple meanings depending on the context. For example, in Italian (which was the language spoken by the study's subjects), the term "Sentire" and its related lemmas can refer to the auditory semantic field ("sento una voce" = "I can hear a voice"), the olfactory one ("Sento un odore" = " I can smell something"), the emotional one ("mi sento male" = "I'm not feeling too good") etc.
If this problem is to be overcome, it is necessary to have recourse to a human disambiguation process. Studies carried out by Fortier and Keen (1999) confirm that the use of informants for studying semantic fields or literary themes is a justifiable enterprise from the statistical perspective.
To this end, the various lemmas were assessed separately by two examiners using the KWIC version of the TACT programmes to ascertain that they truly belonged to the semantic field defined.
TACT (Text Analysis Computing Tool) is a programme which was developed at the University of Toronto for the structural analysis of texts. It had its birth as the result of the need, initially only in the field of literary criticism, for an objective method of studying the works of different authors without being bound by the personal approach of individual researchers.
By means of TACT it is possible to define the "dimensions" of a literary text and obtain information on the contents of the text itself and on its main themes.
For this reason, the application of this specific software to a text allows not only the identification of single individual words but also that of the meaning in that particular context via the statistical processing of the meaningful correlations between different key words.
With TACT it has been possible to obtain the list of KWIC (Key Word In Context) concordances which is made up of lists of symbol-words or key-words (i.e. the most meaningful words amongst those which appear most frequently) inserted in their original context.
The methodology described above permits the underlining of the hierarchical position that symbol-words occupy in dream material by reference to their absolute and relative frequency of occurrence.
The data obtained were analysed using ANOVA

RESULTS
A modest difference in the total number of words used to recount dreams was observed. This difference was not such as to reach a level of significance, however, since it confirmed previous observations made by Watermann.
Significant results were obtained in relation to emotivity. A statistically significant difference between the two groups was observed in the sense that there was a prevalence of explicit statements as to emotional state in the group of younger subjects.
Another highly significant difference between the two groups concerns the uniformity of the verb tenses. When an emotional state or some form of emotional content is explicitly expressed in the elderly subjects' dream texts, a disuniformity of verb tenses occurs and, having begun in one tense (usually the present) the narrative undergoes an abrupt change and turns to the past. This confirms our previous observations (Zanasi, De Persis, unpublished) which demonstrated how explicit statements regarding emotional state have a disruptive influence on temporal uniformity. In other words, one can state that clarification (or lack of it) as to the emotional constituent influences the structuring level or, rather, the consistency of the dream narrative.
The third significant datum regards the semantic fields. A highly significant prevalence in the semantic field linked to the visual sense was observed in the younger subjects in comparison with the elderly ones.

DISCUSSION
The first datum to emerge from a summary analysis of the results and one which is worth emphasising is the presence of significant associations between the descriptive variables examined in their mutual relationship. This appears to confirm the theory of a narrative principle at work in the process of forming the dream which would organise it as "a perceived sequence of events that are connected (a "running" construction) in a non-casual manner" (Toolan's definition of narrative, 1988). In other words, a careful analysis focussing on the dream text's connecting constituents reveals it to be structured in the form of a narrative.
As far as modifications to the expressed emotivity are concerned, these seem to confirm the data in past publications which report a reduced level of explicitly expressed emotivity in the dreams of subjects aged over 50. The fact that the elderly subjects experience a significant "rupture" in the dream text's uniformity of narrative tenses when explicitly expressed emotional states emerge in the dreams can be referred to a greater "vulnerability" on the part of elderly subjects to emotional content. That would lead one to think that Strunz's primitive interpretation linking the lower frequency of emotional content in elderly peoples' dreams to a reduced level of interest should rather be re-read as the expression of a defence mechanism: the dream keeps emotional content at bay on account of its destabilising effect.
Another datum of great interest concerns the difference between the young subjects and the elderly ones in relation to the semantic field linked to vision. This, in our opinion, signals a decline in visual sensory experiences experienced during dreams and reported in the elderly subjects' dream account.
Such a datum cannot easily be interpreted: a possible explanation could be tied to a selective senescence in the neural networks involved in the dream's genesis and those linked to the visuospatial memory, in particular.
Studies using PET (positron emission tomography) (Smith and Jonides., 1997, 1998, Smith et al 1999) show that different neural networks are recruited when different kinds of information need to be stored in working memory. The studies show that different networks are used for visuospatial information and verbal information. Verbal working memory is lateralized to the left hemisphere; it involves frontal regions and posterior parietal regions. In contrast, spatial working memory is relatively bilateral, but with a right-hemisphere dominance; it again involves frontal and parietal areas. The visuospatial working memory undergoes a modification as age advances, in the sense that the right-hemisphere dominance of the frontal regions and posterior parietal regions in young people is substituted, after the age of 70, by a complete bilateralization in the frontal regions, whilst a modest dominance in the posterior ones is maintained. These data would seem to demonstrate that aging leads to less reliance on "specialised" areas and more recruitment of homologous areas in the other hemisphere.
This datum therefore seems to suggest that visuospatial memory becomes increasingly "vulnerable" with advancing age. It is known that in many situations, visual input tends to dominate the other sensory modes of expressing memory and perception and in the fastest responses; visual dominance appears to be related to the relatively weak capacity of visual inputs to alert the organism to their occurrence. In response to this reduced state of alert, the subject tends to keep his/her attention "tuned" to the visual system.
The reduction in vision-associated lemmas could be caused by a reduction in the capacity for active tuning, unlike auditory perception which operates through a more passive reception and does not need to kept tuned. These data confirm the usefulness of our experimental approach to dreams. The technique employed has allowed us to extract from oneiric material a quantity of information on the internal state of the dreamer. This information confirms the Jungian hypothesis of the dream as a "symbolical self-portrait of the internal state of the dreamer." We consider that this method may be used to help deepen our understanding of the mechanisms and purposes of the dream activity as well as having a possible application in the field of clinical diagnosis.

TABLE 1

PLACE	L11     Place designation present. The action takes place in an enclosed environment L12      Place designation present. The action       takes place in the open/outside. L13       Place designation present. The scene in which the action takes place shifts from an open environment to an enclosed one. L20        Place of the action designation absent.
CONTEXT	CON 1  The descriptive constituent is the dominant constituent in defining the context CON 2   The emotional constituent is the dominant constituent in defining the context CON 3   The context is not described
CHRONOLOGY	T1          Action time designation present. T2          Action time designation absent.
NARRATIVE SEQUENCE	SEQ1     Linear narrative sequence SEQ2     Broken narrative sequence
SPEECH	DIS 1    Direct speech used DIS 2    Indirect speech used DIS 3    Alternating direct and indirect speech DIS 4    Scene described from an external viewpoint
CHARACTERS	P1        The dreamer is the only character in the dream scene P2   The dreamer is flanked by "extras" (not characterized) P3       The dreamer is flanked by actors (who take part in the dream text's plot) P4     The dreamer is in the position of an     external observer
SITUATION	S1      Fantastical situation S2      Realistic situation
EMOTIVITY	E1     Emotional state explicitly stated E2     Absence of explicit statement as to emotional state
VERB TENSES	T1   Narrative in the present tense T2   Narrative in the past tense T3  Narrative alternating between present and past tenses
NUMBER OF WORDS	@
SEMANTIC FIELDS	SIGHT HEARING TOUCH SMELL TASTE

Tab. 2
VARIABLE EMOTIVITY

EMOTIVITY
Group	E1	E2	total
controls	88 *	63	151
elderly	62	78	140
total	150	141	291

 * Chi-Square with  1  degrees of freedom  5.9203 P = 0.0150

Tab. 3
CROSSING BETWEEN THE VARIABLES EMOTIVITY AND THE TENSE ONES.

CONTROLS	E1	E2	total
T2x (T1+T2)*	45	43	88
T3	43	20	63
total	88	63	151
ELDERLY	@	@	@	@
T2x (T1+T2)**	27	50	77	@
T3	35 *	28	63	@
total	62	78	140	@

* Chi-square with 1 degrees of freedom= 5.8963 P= 0.0152

(**) This variable  is obtained from the sum of T1 and T2 and represents the homogeneity of the verbal tense in the examined text-dream.

Tab. 4
SEMANTIC FIELDS

(Number of lemmas relating to the senses)

@	CONTROLS	ELDERLY
SIGHT	110 *	72
HEARING	18	11
TOUCH	5	5
SMELL	0	1
TASTE	1	0

* Chi Square with 1 degrees of freedom = 18.59576  p= 0.0001

___________________________________________________________________________________________

REFERENCES

Blick K.A, Howe J.B. (1984), A comparison of the emotional content of dreams recalled by young and elderly women, J Psychol Jan;116 (1st Half): 143-6

Chatman S. (1978). Story and discourse: Narrative structure in fiction and film. Ithaca and London: Cornell University Press.

Danesi M. (1999). Of cigarettes high heels, and other interesting things. New York: St. Martin's Press.)

Domhoff, G. W. (2001). A new neurocognitive theory of dreams. Dreaming, 11, 13-33.

Fortier P.A., Keen K.J., (1999) "The Reliability of Human Disambiguation in Text Markup" ACH- ALLC Conference, Kevin J. Keen and Paul A. Fortier, web resource at http://

Foulkes, D. (1982). Children's Dreams. New York: Wiley.

Foulkes, D. (1983). Cognitive processes during sleep: Evolutionary aspects. In A. Mayes (Ed.), Sleep mechanisms and functions in humans and animals: An evolutionary perspective (pp. 313-337). Wokington UK: Van Nostrand Reinhold.

Foulkes, D. (1993). Data constraints on theorizing about dream function. In A. Moffitt, M. Kramer, & R. Hoffman (Eds.), The Functions of Dreaming (pp. 11-20). Albany, NY: SUNY Press.

Foulkes, D. (1996a). Dream research: 1953-1993. Sleep, 19, 609-624.

Foulkes, D. (1996b). Misrepresentation of sleep-laboratory dream research with children. Perceptual and Motor Skills, 83, 205-206.

Foulkes, D. (1999). Children's dreaming and the development of consciousness. Cambridge: Harvard University Press.

Giambra L. M, Jung R. E., Grodsky A. (1996) Age Changes in Dream Recall in Adulthood Dreaming: Journal of the Association for the Study of Dreams. Vol 6(1) 17-31, Mar.

Gigliozzi G. Il testo e il computer. Manuale di informatica per gli studi letterari B. Mondadori, Milano 1997

Jung C.G. (1945), L'essenza dei Sogni, Opere, Vol.8, Boringhieri, Torino; 1976

Kilroe P. (2000), The Dream as Text, The Dream as Narrative, Dreaming Vol. 10, no. 3

Kramer M., Roth T., Czaya J. (1975), Dream development within a REM period. In P. Levin and W. Koella (Eds.), Sleep 1974: The 2nd European congress on sleep research, Rome,1974 (pp. 406-08). Basel: Karger.

Kramer M., Kinney L., Scharf M. (1983), Dream incorporation and dream function. In W. Koella (Ed.), Sleep 1982: The 6th European congress on sleep research, Zurich, 1982 (pp. 369-71). Basel: Karger.

Kramer M. (1993). Dream translation: An approach to understanding dreams. In G. Delaney (Ed.), New directions in dream interpretation (pp.155-194). Albany: SUNY Press.)

Poster M.I., Nissen, M.I., Klein R.M. (1976) Visual dominance: An information-processing account of ist origins and significance. Psychological Review 83(2), 157-171

Roffwarg, H., Dement W., Muzio J., Fisher C. (1962), "Dream imagery: Relationship to rapid eye movements of sleep. Archives of General Psychiatry, 7, 235-58.

Salvio M.A., Wood J.M., Schwartz J., Eichling P.S. (1992),. Nightmare prevalence in the healthy elderly. Psychol Aging. Jun;7(2):324-5.

Smith, E.E., Jonides, J. (1997).: Working memory: A view from neuroimaging, Cognitive Psychol. 33:5-42

Smith, E.E., Jonides, J., Koeppe R.A., Marshuetz, C.: (1998). Components of verbal working memory: Evidence from neuroimaging, Proc. Natl. Acad. Sci. USA 95:876-882

Smith, E.E., Jonides, J(1999) Storage and executive processes in the frontal lobes, Science 283:1657-1661

Solms M. (2000). Dreaming and REM sleep are controlled by different brain mechanisms. Behavioral and Brain Sciences, 23(6)

Strunz F. (1993) Dreams in the elderly: contents and clinical usefulness, Gesundheitswesen. Nov;55(11):595-601.

Taub, J., Kramer M., Arand D., Jacobs G. (1978), Nightmare dreams and nightmare confabulations. Comprehensive Psychiatry, 19, 285-91

Toolan M. J. (1988). Narrative: A critical linguistic introduction. London and New York: Routledge.,

Waterman D. (1991) Aging and memory for dreams. Perceptual and Motor Skills. Oct;73(2):355-65.

Zanasi M., De Persis C., Analisi Testuale dei Sogni, unpublished

Zadra A. L., Nielsen T.A.; Donderi D.C. (1998), Prevalence of auditory, olfactory, and gustatory experiences in home dreams, Perceptual and MotorSki11s, 87, 819-826.

*Marco Zanasi, M.D.

Neurologist and Psychiatrist, Marco Zanasi is a Jungian analyst and a member of the International Association of Analytical Psychology. He is a Group Analyst and Research Professor at "Tor Vergata" Rome University. For ten years he has studied the correlations between dreams images and psychopathology using Textual Analysis Software.

E- mail : marco.zanasi@uniroma2.it

>>Up