Cultural differences in recalling literary texts

Elly Brosig (Institut für Linguistik/Germanistik Universität Stuttgart)

Summary:

This paper presents the results of a large project investigating the psycholinguistic processes that occur when reading or hearing a meaningful chapter of a literary text. - Since E.F. Bartlett’s famous experiment "The War of the Ghosts" in 1932 very little has been done in this respect. In this experiment Bartlett had shown that the recollection of an American Indian tale was greatly influenced by the different cultural background of his subjects.

For her experiment the author chose an episode from the autobiography of the German author Erich Kästner describing an unhappy Christmas Eve in his childhood. Subjects were 511 students of the author’s seminars, tested in 17 groups, about 60% German nationals, the rest from foreign countries.- The text was read to half of each group; the other half was presented with a written copy for reading. Immediately after this, both groups were instructed to write down as much as they remembered, and in the way they remembered it.

The evaluation substantiated Bartlett’s results.- There were few subjects who did not change anything in the story, in spite of the short interval between input and recall, Germans as well as students from foreign countries, females somewhat more than males. - On closer evaluation it became evident that this was due to an interactive bottom-up, top-down processing .- Separate items in the text, like candles, snow, the smell of cookies - without actually mentioning Christmas - had led them - bottom-up - to the mental schema "Christmas Eve" - and from there top-down to their own personal Christmas memories with very personal details which they added to the story. This was particularly evident in students from Southern and Eastern European countries, who added, for instance, such details as a big Christmas dinner, or the manger with the Infant under the Christmas tree, which featured the different cultural background of these subjects.

These findings seem to reveal much about our processing and recollecting of literary texts. We do not remember exactly what we hear or read, but we adapt a text to our own personal schemata and cultural background and reconstruct it accordingly.

 

Introduction:

The topic of this paper are the psycholinguistic processes that occur in our brain when we are hearing or reading a literary text. Since E.F. Bartlett’s famous experiment with the story "The War of the Ghosts" in 1932, very little has been done in this respect. In this experiment Bartlett had shown that the recollection of this American Indian tale was greatly influenced by the different cultural background of his subjects.

Starting in 1997, the author began a project of literary text recalling with a pilot study and continued it with 17 more groups of subjects until December 2004, in all 511 subjects, all students of the author’s psycholinguistic and sociolinguistic seminars at the University of Stuttgart, Germany.

One of the primary aims of this project had been to investigate possible differences in the remembrance of a heard and a written text presented to each half of the groups, for instance, whether the auditory group remembered more and longer sequences of the text verbatim than the visual group - or vice versa - or whether one of the groups remembered the text rather by propositions.

While this part of the project has not been fully evaluated yet, and the first eight groups not showing significant differences between the auditory and the visual group, the author encountered, from the very beginning, spectacular socio-cultural effects in all groups.

Methods:

For her experiment the author chose an episode from the autobiography of the German author Erich Kästner "When I was a little boy" ("Als ich ein kleiner Junge war") describing an unhappy Christmas Eve in his childhood. A Kästner text was chosen because his writing resembles spoken speech closely, in order to keep the variables for the auditory and the visual groups the same.

The following original text was read to half of each experimental group, while the other half of the group had been sent out:

Ich stand also am Küchenfenster und blickte in die Fenster gegenüber. Hier und dort zündete man schon die Kerzen an. Der Schnee auf der Straße glänzte im Laternenlicht. Weihnachtslieder erklangen. Im Ofen prasselte das Feuer, aber ich fror. Es duftete nach Rosinenstollen,Vanillezucker, und Zitronat. Doch mir war elend zumute. Gleich würde ich lächeln müssen, statt weinen zu dürfen. Und dann hörte ich meine Mutter rufen: "Jetzt kannst du kommen!" Ich ergriff die hübsch eingewickelten Geschenke für die beiden und trat in den Flur. Die Zimmertür stand offen. Der Christbaum strahlte. Vater und Mutter hatten sich links und rechts vom Tisch postiert, jeder neben seine Gaben, als sei das Zimmer samt dem Fest halbiert. "O", sagte ich, "wie schön!" und meinte beide Hälften. Ich hielt mich noch in der Nähe der Tür, so daß mein Versuch, glücklich zu lächeln, unmißverständlich beiden galt. Der Papa, mit der erloschnen Zigarre im Mund, beschmunzelte den firnisblanken Pferdestall. Die Mama blickte triumphierend auf das Gabengebirge zu ihrer Rechten. Wir lächelten zu dritt und überlächelten unsre dreifache Unruhe. Doch ich konnte nicht an der Tür stehen bleiben.
(Erich Kästner, Als ich ein kleiner Junge war)

Translation (by the author):

There I was standing at the kitchen window looking into the windows facing us. Here and there they were already lighting the candles. The snow in the street was glittering in the light of the lanterns. Christmas carols were in the air. The fire was crackling in the stove, but I was cold. There was a scent of raison loaf, vanilla sugar, and candied lemon peel . But I felt miserable. In a moment I would have to smile, instead of being allowed to cry. And then I heard my mother calling: "You may come now!" I took the prettily wrapped presents for both of them and entered the hallway. The door of the living-room stood open. The Christmas tree was radiating. Father and Mother had placed themselves to the left and the right of the table, each beside his presents, as if the room as well as the feast were cut in halves. "Oh", said I, "how beautiful!", and meant both halves. I still kept myself near the door, so that my attempt to smile happily was unmistakably meant for both of them. Daddy , with the extinguished cigar in his mouth, was beaming at the glossily varnished horse stable. Mommy was looking triumphantly at the mountains of gifts on her right. All three of us were smiling, and oversmiling our threefold anxiety. But I could not remain standing at the door.

After the author finished reading the text, the other half of the group was called back in and presented with a written copy of the text, for which they got 2 ½ minutes to read. - Immediately after this, both groups were told to write down as much as they remembered, and in the way they remembered it; that means, not to write a classical literary reproduction. By giving these instructions the author deliberately destroyed the echoic effect of the auditory group, which otherwise might have been an advantage for them.

Results:

The first aim of this project had been to find possible differences between the auditory and the visual group in regard to recalling more and longer sequences of the text verbatim, or rather by propositions. Such verbatim sequences were counted and computed subsequently. While this computation has not been completed for all groups, the difference between the visual and the auditory group after the evaluation of 205 subjects did not turn out to be significant. Moreover, there was a great variance within the groups.

However, the most astonishing - and unexpected - result from the very beginning had been details which subjects mentioned in their reproductions that were not in the story. Among the 511 subjects there were only a few, about 15%, who did not change or add anything to the story, in spite of the short interval between input and recall. - Students from foreign countries behaved the same as Germans in this regard. Females were somewhat more inclined to changes and additions than males, but there were great individual differences within the sexes.

These changes and additions occurred on nearly all sensory levels: visual, auditory, and olfactory, but also on a spatial, cognitive, and emotional level.

The olfactory level proved to be the strongest one. The subjects reported smells, for instance, roasted apples, spiced cookies, and marzipan, all characteristic of a German Christmas - but did not pertain to the story. The most frequently named scent was cinnamon.- However, already in the first experiment the author encountered scents very untypical of a German Christmas, and she realized that these might reflect cultural differences with regard to Christmas customs.

At this point, the author began to ask for the nationality of the subjects. - A Spanish male and a male from the former Yugoslavia both smelled lemon pie, which probably is a special dish on Christmas for them; two Greek girls smelled raisin and vanilla bars.- A girl from Georgia smelled scented candles, while coriander and rosemary were named by subjects who did not reveal their nationality, but were not German (which was evident by the mistakes they made in their reproduction). I also could not make out the nationality of the girl who smelled camellia blossoms. - Not all subjects smelled sweet scents, but a German, as well as a Russian, and a French girl smelled the enticing scent of roast meat from the oven.

On the visual level subjects changed the scene of the story. While in Germany the supper on Christmas Eve is rather simple in most families - there are lots of cookies and sweets to eat - it became evident that in other countries it is regarded as a very important meal. - A French girl stated that Father and Mother were sitting at the table, which was beautifully laid; another French girl mentioned that the table was laid, with the forks on the left side. (!) - Two Polish girls also mentioned the beautifully laid table, and Father and Mother already sitting there.- A Chinese girl wrote that the Christmas dinner was ready, and a male subject of unidentified nationality, but not German, stated that there were three different things to eat for dinner. -

Typically none of the afore-mentioned foreign subjects mentioned the Christmas tree, which , of course, is not the custom in all cultures. - While stating that the room was beautifully decorated, neither a Russian, nor a Polish girl mentioned the Christmas tree either.-

Obviously, all these diverging details occurred on a visual level, as did the manger with the Infant under the Christmas tree, which a number of German subjects, but also an Italian imagined. - While the author did not ask for their religious confession, she is certain that these subjects were Catholic, because in Germany you will rarely find a manger in Protestant families. - Another very frequently named object was a fireplace - which, of course, the rather poor Kästner family did not own.

On the auditory level, the most frequently mentioned detail, by German subjects only, was the little bell ringing, which calls the children to the presentation of gifts in the Christmas room. - A non-German male - the same with the three different dishes for dinner - stated that the mother was reading poems from a book. - A German girl wrote that the family was wishing each other "Fröhliche Weihnachten!"(Merry Christmas). Another stated "We were singing Christmas carols", and another heard someone singing "Oh du fröhliche!", all definitely German Christmas customs - but not in the story.

Apart from the sensory levels, there also appeared changes on a spatial level. Several subjects changed the setting of the scene: Instead of in the kitchen, the child was waiting in his own room upstairs and had to go downstairs to the Christmas room. This is a clear evidence that they imagined their own home. - This may also explain the frequently mentioned fireplace. But, of course, there may also be the influence of formerly read or heard Christmas stories, which had become part of their schema.

Some of the deviations from the story might have originated on the cognitive level, with subjects rationalizing what might have been, when they did not remember it. But even in this case, the individual experience - their personal schema - will have interfered.

But there were also subjects who changed the emotional level of the story. - Most subjects understood and recalled that the child was unhappy because of his parents’ rivalry for his love. But some subjects - more males than females, and all German but one - further dramatized the situation by stating that the parents had separated, or were even divorced. From the way they expressed this - sometimes even with exclamation marks - the author got the impression that they projected their own family situation into the story, and were themselves unable to enjoy Christmas. Some male subjects even expressed their hostility toward the whole Christmas situation.

On the other hand, there were also subjects - among them a Chinese girl - who seemed so overwhelmed by their own happy Christmas memories that they wrote how beautiful everything was, how they exchanged gifts with their parents, and that everybody was happy.

Interpretation:

When it became evident that the afore-mentioned astonishing phenomena were not singular incidents, but occurred in nearly every reproduction, the author began to seek a scientific explanation. - She got the clue when she encountered the first reproduction starting with the sentence "It was Christmas Eve" (Es war Heiligabend). (In fact, a number of reproductions started with this sentence.) This was a definite indication of bottom-up, top-down processing.

This bottom-up, top-down processing is part of the schema theory by Rumelhart (1980). He assumes that our world knowledge is organized in a neural network of schemata (other authors call them scripts or frames), i.e. bundles of information about a topic, like riding in a bus, a visit to a restaurant - or Christmas Eve. These schemata are partly general, and partly individual knowledge (semantic and episodic memory).

This model of language processing functions bottom-up (data- or sensory-driven) from the smallest level (the phonemes) upwards to the highest level (the schema), i.e. from particular details to the specific schema. - From there the processes run top-down (knowledge-driven), from the specific schema to particular details. These processes are working parallel, i.e. interactive. Thus, it is an on-line interactice model of language processing, which is a connectionist model, as described by McClelland and Rumelhart,1986, (a good survey in Gleason/Ratner 1998).

This interactive activation model is compatible with our knowledge of the brain. A stimulus, for instance, a certain word, activates specific neurons, which in turn spread their activation via their synapses to other neurons with which they have previously been connected on other occasions. Frequently activated neurons remain connected in assemblies, and are more easily activated by repeated input of the same stimuli. The whole pattern can also become activated, if only some part of it is exposed to stimuli. This can explain bottom-up (from words to a schema), as well as the top-down processes (from the schema to individual details connected with it).

So what evidently happened to the subjects was that details mentioned at the beginning of the text, like candles being lit, snow in the street, the smell of typical Christmas spices - without actually mentioning Christmas - had led them bottom-up to their schema "Christmas Eve" (in Germany an emotionally very strong schema), and from there top-down to their own personal Christmas memories, including the details pertaining to it.

The author’s first experiment had been carried out on a cold winter day, shortly before Christmas, and at first she supposed that the results might have been influenced by that. So next time she repeated the experiment on a very hot summer day in an unclimatized classroom.

The results remained the same! - So it seems that our schemata are really firmly anchored in our brain and are not easily changed.

At the end of her evaluations, the author compared her results to those of E.F. Bartlett’s "War of the Ghosts", mentioned in the beginning. - It has to be noted, however, that Bartlett’s experiments were in some respect different: He concentrated on changes and additions that occurred during repeated reproductions of the story, and he attributed the changes in details or in the story to processes occurring in long-term memory. - But while on the whole our basic findings were much alike, it became evident from my own experiments that such changes occur already in working memory (STM) right after the input.

Aside from that, there is a lot of agreement. For example, Bartlett states in the summary of his studies:

"It appears that accuracy of reproduction, in the literal sense, is the rare exception and not the rule...... The actual memory process is strongly and evidently constructive, and there is much use of inference..... Details are remembered when they fit with a subject’s pre-formed interests and tendencies and are often transformed accordingly..... Names, phrases, and events are immediately changed so that they appear in forms current within the social group to which the subject belongs."

Bartlett also stressed the affective attitude of the subject, his emotional state, that influences how, and what part of the story will be remembered. - He also states that

"rationalization may deal with details linking them together and so rendering them apparently coherent, or linking a given detail with another detail not actually present in the original setting."

The author can substantiate all these findings from the results of her experiment. - However,

nowadays after 70 years’ progress in psycholinguistic research, we can interpret these results further with the interactive bottom-up, top-down model, which is a connectionist model.

Of course, there have been other authors trying to explain text processing since Bartlett. Some of their models have a connection to the author’s results, for instance, van Dyck and Kintsch (1983) who assume three levels of internal representation in the comprehension process:

1. a verbatim representation, which is derived directly from the surface of the text:
2. a text base representation, which is based on the concepts and meaning of the text;
3. a situation model,, which is a combination of prior knowledge, including the

goals and attitudes of the reader, and text information.

The verbatim and text base representation contain purely text-based information (data-driven), whereas the situation model contains information how the reader interprets the text (top-down processing).

Another model that comes close to the author’s findings is from Bransford et al.(1972), who postulate:

1. Mental representation of a text is not a representation of its linguistic form.
2. Text comprehension is an integrative process.
3. Text comprehension is a constructive process. Information that is explicit in a text has to be combined with relevant information about the world.

This is basically an interactive bottom-up, top-down model.- These and other models are presented in Britton/Graesser (1996), J.M.Mandler (1984), and R.J. Spiro (1980).

The greatest problem with all of these models is, however, that while nearly all of them have elaborate theories of text processing, their experiments trying to prove them are not realistic. - Nearly all their studies are very small; some limit their subjects to pre-school children; others have a merely descriptive text recalled sentence by sentence. Results that seem irrelevant are not taken into account and are suppressed. None of the authors let their subjects recall an authentic literary text, as did Bartlett. Of course, they also had partly different intentions with their projects.

Moreover, most models do not take into account that the process of text comprehension occurs inside of individual persons with very individually structured memories about the world they encountered so far. But these have a bearing on texts that are more fictional than descriptive. But even in these cases an irrelevant top-down processing can be found. - Who has not encountered in examination papers the phenomenon that students let themselves be carried far from the topic in question, even if the question had been quite precise? - That means that even in strictly scientific texts, subjects are influenced by their own social attitudes and background.

Conclusions:

The results of this project seem to reveal much about our processing and memorizing literary texts. We do not remember what we hear or read, but we adapt it to our own personal schemata, and reconstruct it by interactive bottom-up, top-down processing. - The advantage of this connectionist model is that it can be explained on a neuro-physiological basis, which makes it more modern and current.

© Elly Brosig (Institut für Linguistik/Germanistik Universität Stuttgart)

REFERENCES

- Bartlett, E.F. 1932 Remembering (Re-issued 1995) Cambridge, Mass.: University Press

- Bruce K.Britton/Arthur C. Graesser (eds.) 1996, Models of Understanding Text. Mahwah, N.J.: Lawrence Erlbaum

- Jean B. Gleason/Nan B. Ratner (eds.) 1998, Psycholinguistics, 2d ed., Fort Worth: Harcourt

- Kästner, Erich 1957, Als ich ein kleiner Junge war. Zürich: Atrium Verlag

- Mandler, J.M. 1984. Stories, scripts and scenes. Aspects of a schema theory. Hillsdale: Lawrence Erlbaum

- McClelland J.L./Rumelhart, D.E. & PDP Research Group 1986, Parallel distributed processing: Explorations in the microstructure of cognition. Vol.2: Psychological and biological models. Cambridge, MA: Bradford, MIT Press.

- Rumelhart, David E. 1980, Schemata the building blocks of cognition, in: R.J. Spiro, B.C. Bruce & W.F. Brewer (Eds.) Theoretical issues in reading comprehension. Hillsdale, NJ.: Lawrence Erlbaum.

- Spiro, R.J. 1980, Constructive Processes in Prose Comprehension and Recall, in: R.J.Spiro, B.C.Bruce & W.F. Brewer (Eds.). Theoretical issues in reading comprehension. Hillsdale, NJ.: Lawrence Erlbaum.

6.4. Innovations in Psycholinguistics: A Step to Innovations in Brain, Culture, Cognition and Communication Research

Sektionsgruppen | Section Groups | Groupes de sections

       Inhalt | Table of Contents | Contenu  16 Nr.

Webmeister: Peter R. Horn     last change: 27.4.2006