"Where was I?": Personal Experience Narrative, Crystallization and Some Thoughts on Tradition Memory¹

Several years ago, I was rather surprised when one of my daughter's friend's parents asked me if I could help her with brain surgery in a few weeks' time. She assured me that she was, in fact, a medical doctor and a practicing surgeon and that the request was sincere. We would not be performing amateur neurosurgery on an unsuspecting passer-by as part of a bizarre Southern Californian cult ritual. Rather, she would be trying to help a young woman regain her health by excising a remarkably aggressive tumor lodged deep in her brain. I indicated that two weeks might not be enough time for me to learn how to perform surgery, particularly something as delicate as neurosurgery, as opposed to something coarser, say an appendectomy or a simple amputation. I also reminded her that my doctor title was strictly related to philosophy: "Unless your patient is suffering from the effects of repeated attempts at deconstructing literary texts," I said, "or an inexplicable desire to perform Finnish epic songs while strumming a kantele, I doubt that I can be of much use." But my "colleague" persisted, and explained that she needed me more for my Danish language skills than any alleged ability to decode literary texts or analyze traditional expressions. Having only encountered one other situation where I was sought out for my ability to speak Danish (that involving Kelsey Grammar, a Snickers bar, and a large multinational telecommunications firm), I acquiesced. This opportunity also allowed me to enter something on my shared university calendar that, if encountered by nosy legislators, might finally justify my employment.

The surgery was relatively straightforward from my perspective. The cranium had already been removed by the time I arrived in the operating theater, and no one asked me to close at the end. My task was simple: I was to speak Danish with the very conscious patient, making sure that she recognized pictures on flash cards, could maintain as coherent a conversation as possible while someone rooted about in her brain, and that she did not begin to speak nonsense or make strange noises. The patient was a young woman whose first language was Danish but who had learned English at such an early age that she was considered to be a "true" bilingual. The goal, as my colleague put it, was to avoid "cutting out" her Danish; they had "mapped" her language centers but, because she was bilingual, the fear was that her Danish language did not map to the same area as her English. The neurologist explained that language for this patient could have more than the one center that is common in most monolingual patients. By applying electric currents to various parts of the brain while the patient spoke, my colleague could check the preexisting map and avoid paths toward the lesion that would disrupt possibly fragile language networks. The surgery was a success, the lesion was removed, and, the next day, I had a wrap up conversation with the sedated but grateful patient. Her Danish was fine. My colleague had said she was worried about the odd sounds coming from the woman during surgery, but I assured her that it was standard Danish.

While the patient's linguistic capacities remained intact, the experience brought home to me the very physicality of language and, by extension, memory. Often, in the humanities, we conceive of human expression—language and the expressions that are built around language—in ways that are quite divorced from this physicality. But my experiences during this surgery made it quite clear—as clear as it could possibly be—that language and memory (and, by extension, learning) are, among other things, connected to physiologic structures and the result of neurobiological processes. The pre-surgical preparations and the surgery itself also made absolutely explicit that no two brains are alike (Schumann 2004). Not only does each individual have a brain that has been shaped by their genetic inheritance, but that brain has been further shaped by (and is constantly being shaped by) environmental factors—either inputs through the senses or very real physical changes caused by disease, injury or, in this case, deliberate intervention (Schumann, Crowell, et al. 2004). Several weeks later, at a social gathering, I met my colleague and her husband—also a neurosurgeon—and I began to ask questions about the implications of the physiological structure of the brain and the neurobiological processes linked to language, learning, and memory that might help us understand traditional expression in ways other than those prevalent in the academy. Since I was so flabbergasted by the idea that language exists discretely in a section of the brain so readily identifiable, I wondered if tradition too might be linked to special mechanisms—or even specific sites—in the brain that differentiate tradition from other forms of memory.

In short, the surgery sparked my interest in the neurophysiologic processes of tradition and touched off a series of questions that I will attempt to address, however inadequately, in this short essay. Is there such a thing as "tradition memory"? Do traditions present in neurological terms differently than other types of memory? Does the process of expressing traditions verbally present in a neurologically different manner than what cognitive psychologists label "normal speech"? Is there a neurobiological process behind Hymes's (1975) breakthrough into performance? Can an appreciation of neurobiological processes also help explain the long apprenticeship and subsequent mastery of epic tradition among the singers encountered by Milman Parry and Albert Lord (Lord 1964)? Does one learn and process tradition as language and, if so, is it more akin to native language acquisition or second language acquisition? Or does one learn and process tradition as lived experience? How does one transform lived experience into traditional expression? To what degree is tradition production—acquisition, storage, and, most importantly, retrieval—automated compared to natural native language or other habitual skills? Or is it not an automatic process at all? Is it an automatic process for some people and not an automatic process for others? In other words, to what degree is tradition conditioned by declarative memory and to what extent is it subsumed by procedural memory? Similarly, why are some people so attracted to aspects of a tradition (active participants) that they master it and others do not (von Sydow 1948)?² What accounts for the varying levels of skill and attraction that one finds in any community in regards to all of the community's traditions? Speaking more basically, to what extent is tradition conditioned by neurobiology and genetic inheritance, and to what extent is it conditioned by environment? This essay is intended to begin asking these questions about tradition in a most preliminary fashion while appealing to theories of memory, learning, and forgetting. The tentative suggestions put forward here—and I stress tentative—might help us widen our ability to understand how the dialectic tension between individual and tradition that is the fundamental basis for folklore functions not only in society but, quite physically, in individuals who make up those societies (Chesnutt 1999).

Before I speculate briefly on these questions in the context of a small set of folkloric data (a chain transmission of the legend "The Hook" and a network transmission of "The Fifty Dollar Porsche" both by a group of UCLA undergraduates), it seems prudent to provide an overview of current models of memory in the brain and to describe, however briefly, some of the main approaches to memory and tradition. I will conclude with a short overview of Multiple Trace Theory (MTT), a theory that has some significant advantages over the Standard Theory of memory consolidation in the context of understanding variation and stability in tradition, and can perhaps provide some more insight into those small experimental data sets. The two small data sets have significant constraints as suitable test data, and it may well be possible to design more rigorous experiments to test these hypotheses in the future.

Questions of variation and stability in tradition have been quite vexing ones for folklorists, at the same time as those two features have also helped define the bounds of the field itself. Early folklorists often neglected to pay attention to the actual individual tradition participants, positing a superorganic view of tradition that was very far from a view of tradition that could incorporate considerations of individual brain structure and function (Krohn 1926). Even Walter Anderson's (1923) classic theory of "self correction"—a notion that a story would be brought back into line with its traditional form by other tradition participants because of repeated telling of the story and repeated hearings of the story through time—suggests that stories have a life of their own. In its best articulation the "law of self correction" is related to questions of memory and reinforcement—a person who has heard and remembered a story multiple times and from multiple sources is likely to reject external idiosyncrasies in favor of his or her own memories. More recent studies of individual repertoires and world view, such as those of Pentikaïnen (1978), Siikala (1990), Kaivola-Bregenhøj (1996), Dégh (1989; 1995), and Palkó and Dégh (1995), as well as my own study of nineteenth-century Danish storytellers (Tangherlini 1994a), place far more emphasis on the individual and the complex societies in which he or she lives and participates in traditional practices. Yet, in all of these studies, there is no real discussion of why individuals might have different repertoires. Generally, these studies simply offer the observation that these differences exist and reflect an aspect of an individual's world view (Pentikäinen 1978).

A decade or so ago, David Rubin (1995), in his Memory in Oral Tradition, attempted to bring the perspectives of a cognitive psychologist to bear on a narrow range of oral traditional genres. In his study, he concentrated on rhymed or sung (or rhymed and sung) traditions, and attempted to align findings from folklore theory—particularly Parry and Lord's Oral-Formulaic Theory—with the then current theories about remembering and forgetting. He proposed early on in his study that learning traditional expression—particularly genres such as counting out rhymes, ballad singing, and epic singing—is equivalent to learning one's native language, stating, "genres of oral traditions can be considered as poetic languages, or overlay systems, or rule bound registers of speech. Second, learning such poetic languages is similar in many ways to learning first languages" and offered five points to support this hypothesis, including the important—yet potentially false—observation that "the process of learning and what is learned are not available to introspection" (136). In this context, "introspection" means that an individual is not only aware that he or she is learning but is also aware of what, and how, he or she is learning. Rubin opined that, while "Memory is often considered as a storehouse…for oral traditions a better metaphor is that of a well-practiced skill dependent on extensive experience" (Rubin 1995, 146). This latter observation proposes that tradition is largely related to automatic processes of nondeclarative, procedural memory, a position that is somewhat extreme (Rubin 1995, 136). Not all traditional performance is automatic, and a great deal of oral tradition is available to introspection, particularly for the most competent and active participants in that tradition.

The underlying theoretical orientation of Rubin's (1995) approach to aspects of stability and variation is described as "cue-item discriminability." This position holds that recall [of oral traditions] starts with the first word of the song and proceeds in a linear fashion. Words sung are cues for words yet to be sung. If words are to be recalled, they must be discriminated from other words in memory. The general constraints of the genre and piece, especially rhythm, act as cues from the start, with the singing filling in other cues as it progresses. This process, after the initial, often conscious decision to sing a song has been made, can go on without conscious intervention, using what has been called implicit (nondeclarative) or indirect memory. The serial-recall method, however, means that knowledge in oral traditions is not routinely accessed without the cues provided by a running start and often cannot be accessed without them. (192)

Thus, the performance of a traditional expression, once begun, proceeds automatically, with little recourse to introspection. Rubin, of course, has overstated the case for sung tradition in his emphasis on the "running start." Although many tradition participants often need a "running start" to access certain expressions, others—those that Bengt Holbek (1987) characterizes as the "craftsmen" of tradition—do not.

The example that Rubin offers of the folk singer who needs to wait for the chorus to come around again before he can sing it is so recognizably incorrect for "craftsmen" of a tradition that the folk singer Arlo Guthrie (1967) includes a humorous gloss on it in his well-known live rendition of "Alice's Restaurant." In that song he addresses his audience, saying, "So we'll wait for it to come around on the guitar here and sing it when it does. Here it comes…" In response, the audience laughs, acknowledging the unlikely premise that Guthrie needs this type of cue to sing the refrain even though they might well need that type of cue. While it is probably true that, for most tradition participants, this type of cue-item discriminability is necessary for successful recall of a traditional expression, for the most active and competent tradition participants, access to any part of the expression is far less dependent on cueing. Studies of singers of epic, such as those by Lord, confirm the ability of expert singers to begin and end singing in multiple points in the epic, and to modify their singing quite dramatically to respond to the exigencies of the immediate performance context (Lord, Mitchell, et al. 2000). Of course, on the opposite extreme, the least active tradition participants may not be able to produce a recognizable variant of a traditional expression irrespective of the number of cues they are given.

Rubin's theory also does little to explain features of non-sung, non-rhythmic tradition, since many of the cues to which he refers are dependent on formal features of rhyme or meter. Rubin presents memory essentially as a mysterious, black-box phenomenon and does not explore the actual processes by which memories are consolidated, stored, and activated. A more contemporary model of memory might help explain the "running start" that he situates as a fundamental component of traditional recall and, at the same time, explain in a more nuanced manner how it is that the "craftsmen" of tradition—the most expert of active tradition participants—might be able to eliminate the need for the "running start." Such a model might also explain how processes that for some are "automatic," and not available for introspection or other types of modification, are much more easily accessible to other, more active, tradition participants, both for introspection, innovation, improvisation, and reformulation.

For most tradition participants, traditional expressions—their memory, their recall and to a certain extent, their performance—are not purely automatic processes as Rubin implies. Years of ethnographic and folkloric research has revealed that individuals' participation in tradition is best defined on a scale in relation to the least active to the most active participants in a tradition (fig. 1). It is likely that the automatic processes Rubin describes are a fitting characterization of the manner in which tradition participants who cluster toward the middle of that spectrum remember and perform tradition. For other tradition participants, cue-item discriminability is no longer as necessary—for example in the case of tradition participants who have mastered the particular expressive form—or not functional—for example in the case of individuals who have yet to learn or participate in the tradition enough to be able to remember and reproduce expressions (fig. 1).

In this illustration, the cue-item discriminability Rubin describes is only functional for those tradition participants who are active in the tradition, yet not masters of it. For masters of a tradition—such as Arlo Guthrie in the example above—the song (in this case) is not only readily available for introspection but also for deliberate modification (which is different from accidental modification that one would find at the other end of the spectrum). That is not to say that cue-item discriminability does not play some role in their recall; rather, these "craftsmen" have many more pathways to begin or restart an expression and many more pathways to connect or move between traditional expressions. In contrast, those just learning a tradition have not consolidated the formal features or content to make the type of automatic recall described by Rubin possible. Given the range of variation one encounters in traditional expressive performances, from the performances of the highly competent to the borderline incompetent, there must also be varying degrees of physical connections in the minds of these people for any given expression to explicit or declarative memory.

If we are to advance our understanding of how tradition functions on the individual level, we must move beyond the "black box" approach to learning and memory that is implicit in most studies of tradition, and incorporate in our studies—at least on some level—a consideration of both the physiology and the neurobiology of memory and learning. Schumann points out that "Psychological theory almost universally assumes that across individuals brain structure is homogeneous. Thus, most psychological research on learning proceeds on the notion that all brains are the same…from the perspective of neurobiology, brains are as different as faces" (Schumann, Crowell, et al. 2004, 2). This awareness, that different people learn and remember not only different things but also do so differently may not tell us too much about why traditions persist, but it may help us to understand how traditions persist.

The question of why traditions persist might be subsumed under the neurobiology of motivation and aversion—rewards and value structures can be conditioned by the group and result in measurable changes in neurobiology. These changes result in people seeking out or rejecting environments that, in turn, lead both to learning and to underlying changes in the individual's brain structure. In a continuous feedback loop, these changes again strengthen the motivation and aversion circuits. What Schumann points out for the context of learning can be extended to learning traditions:

The first interaction involves passive effects, in which the parents, whose genes the child inherits, provide the major environmental input to the child. In the second kind of interaction, the individual chooses and/or creates environments that are compatible with his or her talents. The third way genotype interacts with the environment is through the evocation of responses from the environment. Evocative effects are those that an individual elicits from others. (Schumann, Crowell, et al. 2004, 16)

As a result, "interindividual variations are not seen as exceptions or noise, as in traditional psychology, but rather they are considered as a universal basis on which theories of human cognition must be built" (Schumann, Crowell, et al. 2004, 18). These considerations move us significantly away from an approach that considers folklore simply as behavior (Georges and Jones 1995). Instead, it allows us to explore folklore as part of a complex interaction between the individual and his or her physical and cultural environment as well as the individuated processes and effects related to learning and memory.

Memory is generally broken into two main categories: short-term or working memory, and long-term memory. Schumann notes, "Working memory has traditionally been defined as memory that is held for short periods of time (less than 20 seconds) in order to achieve success at a task…Long term memories are those lasting for extended periods of time, from days, to weeks or for as long as months or years" (Schumann, Crowell, et al. 2004, 4-5). As implied earlier, long-term memory is, in turn, broken into two main categories: declarative or explicit memory and nondeclarative or implicit memory. Again, from Schumann's summary of a memory taxonomy, one learns that "declarative memories are memories for facts and events, and nondeclarative memories are memories for habits, motor and perceptual skills, and emotional learning" (Schumann, Crowell, et al. 2004, 5). Each of these two main sets of long-term memory is further broken down into subsets. The subsets of declarative memory are semantic memory and episodic memory. The subsets of nondeclarative memory are conditioning, priming, and procedural memory. It is the last subcategories of each main category of long-term memory—episodic memory and procedural memory—that are of greatest interest to students of tradition. The interaction between these two types of memory—and indeed a particular type of interaction between these two—may well be the locus of what could be called "tradition memory."

Declarative memory—including the important subset of episodic memory that is implicated in storytelling and traditional expression in general—is centered in the hippocampus. What is interesting is that, once stored, declarative memories are not static. Rather, "memories that have been previously stored have already modified the brain in such a way as to affect the relative ease with which new memories can be formed. In other words, learning not only results in memory but is itself the result of memory…at the cellular level, encoding, storage and retrieval are represented as modifications in the strength of synaptic connections that are constantly being altered as the result of new interactions with our environment" (Schumann, Crowell, et al. 2004, 76). Adding to this "dynamic nature" of long-term, declarative memory that once was considered to be remarkably stable, is the recent discovery that "episodic memories may maintain traces in the hippocampus for an indefinite period," a significant departure from "the traditional model, articulated by Squire (Squire, Knowlton, & Musen 1993), [that] suggests that consolidated memories are eventually stored in cortical circuits that are independent of the hippocampus" (Schumann, Crowell, et al. 2004, 95). This dynamic view of declarative memory is an important shift in understanding the physiology of memory and recall and has significant implications for an understanding of stability and variation. Similarly, the interaction between procedural memory and declarative memory as part of the neurobiology of memory consolidation should inform this understanding.

There are two main theories about the consolidation of long-term memories and the manner in which those memories are recalled. Both of these models rely on the relationship between declarative memory and nondeclarative memory. The first, known simply as the Standard Theory of memory consolidation, was first proposed by Müller and Pilzecker (1900). This theory suggests that memory consolidation is time dependent—the memories are first formed in the hippocampus and transferred over time to the neocortex. The traces between the hippocampus and the precortical regions eventually weaken and disappear, while the connections within the neocortex strengthen through repeated recall. Consequently, the memory becomes both stable and less susceptible to disruption or damage in its stable state in the precortical regions. The retrieval of these memories relies on a single—or a small number of—indexical links. This type of indexing explains the "running start," or cue-item discriminability, emphasized by Rubin, as well as his emphasis on the seeming stability—and automaticization—of traditional expression.

This Standard Theory has been supplanted in recent years by Multiple Trace Theory (MTT). Based on evidence that older consolidated memories can, in fact, be subject to disruption of a kind nearly precluded by the Standard Theory, MTT proposes that the connection between the hippocampus and the neocortex do not dissipate, but rather are continuously reinforced, and new connections from the hippocampus to the neocortex indexing the same memory—and other related memories—are constantly being formed. This dynamic model for long-term memory consolidation aligns better with observations in both stability and variation in tradition (Tangherlini 2003) and also provides a mechanism that contests the notion of serial recall—a hypothesis that is easily falsified by both fieldwork and archival data.

One of the key findings of Nadel and Moscovitch who first proposed MTT in 1997 is summarized by Nancy Jones as follows:

As memories are retrieved and rehearsed, multiple traces are made in the hippocampus…These traces are indexed to locations in the neocortex. Each time a new set of hippocampal traces is made they are also indexed to the cortex. Thus, each time a memory is rehearsed, previously linked cortical regions would be linked to another set of traces. Additionally, as more associations are made, new cortical regions could be added to the total set of traces for the given memory. (Schumann, Crowell, et al. 2004, 117)

Linking of this type allows for the possibility of stories growing and shrinking depending both on performance context and the narrator's own development as a person. At the same time as it helps explain elasticity, it also helps explain aspects of stability over time—the "underlying" memory persists, it just has multiple indices in the hippocampus that form, are reinforced, or disappear over time.³

A key feature of MTT that differentiates it from Standard Theory is the introduction of the possibility of disruption of older, consolidated memories. In MTT it is, "not only temporal duration but also the state of the memory (i.e., whether it has been activated or not) that can affect the stability of the memory" (Schumann, Crowell, et al. 2004, 120). This observation helps explain, for example, the wide range of variation one finds between variants of a story told by an active participant—whose memories might be frequently activated, and thus show both a high degree of stability, and a significant amount of indexicality between the hippocampus and the neocortex, thereby allowing for multiple "ins" to the storytelling—and a passive tradition participant whose memory of that same story may have been disrupted either by other environmental inputs, or by the lack of frequent activation, or both. At best, this passive tradition participant would need the "running start" to activate the initial index to the disrupted and weakly linked memory in the neocortex. Indeed, he or she may need multiple running starts or additional prompting to access these memories—a phenomenon attested to by many fieldworkers.

The fragility of long-term memories—including long-term, episodic memories—is addressed well by MTT. In various studies such as those by Loftus, Miller, and Burns (1978) episodic memories for events were shown to be alterable by post-event information—a situation that mimics well the impact that tradition can have on individuals in the creation of personal experience narratives. I would argue that the paramedics with whom I worked often had their episodic memories for events "interrupted" by post-event information, often provided during the initial telling of the newly forming personal experience narrative (Tangherlini 1998). These "interruptions" were, of course, part of the tradition itself, and helped the medic consolidate the experiences of the events into a narrative episodic memory that conformed to the expectations of the group yet maintained the unique aspects of the discrete event (Tangherlini 1998; 2000).⁴

The connection between episodic memory and procedural memory is equally important in understanding the neurobiology of tradition. Rubin (1995) rightfully points out that certain aspects of traditional performance seem almost automatic. This observation has been more formally expressed by Hymes (1975) in his consideration of "breakthrough into performance," an observation that has also conditioned an entire generation of folklorists to focus almost exclusively on aspects of performance. Clearly, there is a degree of nondeclarative—primarily procedural—memory that informs the performance of tradition which, in turn, is based largely on the consolidation and recall of declarative, episodic memories. Jones notes that "Ullman et al (1997) presents a dual model for language in which they posit that the lexicon is processed by the declarative memory system and grammar is processed by the procedural memory system" (Schumann, Crowell, et al. 2004, 124). By analogy, one might suggest that the episodic memory of events or learned events (e.g., narratives) are processed by declarative memory and the performance of those memories—including aspects of genre—are processed by the procedural memory system. Lee notes that "one acquires…[procedural] memory…through the repeated execution of a task…[it] is used for example when one learns how to play a musical instrument, how to dance, how to play a sport, or how to speak native language," and to this list one could add how to tell a story, sing an epic, perform a jump rope rhyme and so on (Schumann, Crowell, et al. 2004, 44). The repeated execution of the task may take the form of both listening and performing; this learning process consequently engages both declarative and nondeclarative memory. In recall, the more stable nondeclarative memories structure the performance, while the more easily disrupted—yet potentially quite stable—declarative, episodic memories provide the content for the performance. Ultimately this approach allows for a holistic understanding of the neurobiology of tradition, allowing for both variation and stability not only in the memory of individuals but also in traditional performances across individuals within a tradition group.

A chain transmission of a single story stands as an apt illustration of the potential for instability in tradition, particularly when members of the chain may be either inactive tradition participants, or completely unaware of the tradition (Anderson 1951). Unlike a regular tradition group, a chain transmission does not allow for the repeated execution of the task—neither telling nor listening. As a result, the memories created during the chain are likely to be more fragile and less likely to reflect the stability that is a hallmark of tradition. In a brief experiment, I asked twelve students in a folklore class to tell, in chain fashion, a version of "The Hook." I told the story to the first person in the chain as follows:

C₀: I'm going to tell you a story called "The Hook." Now this happened when I was a kid and I grew up in central Massachusetts in Worcester and just outside of Worcester. But I remember in high school, this happened to a couple of, of friends of mine. They had gone to a party and after the party they drove out to a lake that everybody would go to after parties, particularly couples, its called Lake Chagoggagogchagoggagogchabunagungamog. And it was outside of Worcester, and it was called Lake Webster. They had parked by the lake and they were sitting there discussing homework. They had the radio on, a little Meatloaf song was playing and they were really getting into their discussion. Very hot and heavy in this discussion. He was very excited, they were going to get to a part of the homework, that he really thought was going to be great, but just then on the radio, a voice broke in and said that an inmate had escaped from Worcester State hospital which was the insane asylum in Worcester. And the way that you could recognize that this was the escaped inmate was that instead of a right hand, he had a hook, and so if you saw this man wandering about you should be very careful because what he liked to do, he liked to slash people with this hook that he had, and so the girl started getting very agitated, and said we have to go home, we have to go home, and he said, "No, no, no, we were just getting to the best part." And she said, "No, no, no, I can't stay out here, its too deserted out here, by the lake and I insist, you have to go home." So finally after some back and forth, the boyfriend got very angry at her, and started up the car, and peeled out of the parking lot. Just you know burned a whole lot of rubber and headed off home to Worcester. And they didn't talk the whole way home. But finally they go to her house and sort of as a gesture of chivalry, he decided to get out of the car, sort of as, you know, an ironic gesture, to show how much of a gentleman he was, he pulls up to the front of her house, and he gets out of the car, and walks around the car to open her door, and there, on the car handle is the bloody hook. So that's the story.

Even though students who self-identified as eager storytellers were deliberately chosen to be the early links in the chain, after two links, the story had already begun to change significantly. Details fell away, and, more importantly, the motivation of the two young lovers to visit the lake all but disappeared:

C₃: Basically, there's this guy and he's in college and in Worcester, Boston, oh no, near Boston, Worcester? And he really likes this girl, and him and this girl they and park somewhere, in front of Lake Webster, which also has this really long complicated Indian name, but it's hard to pronounce, like Chamackamunga or something like that. And so, I don't remember it, and uh, anyway, so they're like sitting in the car and having this crazy conversation like talking about something that happened in class the other day and he's like really into it and everything like that, and I think there's some song from a band on the radio who they both like, or whatever, I don't remember the band name. I don't know who the band name is. Um, and all of a sudden, there's this interruption on the radio of this like this emergency announcement that there's crazed, deranged escaped mental patient in the area, right in the area they were in, and you know anyone in this area should leave immediately, and the only way you can identify this guy is he has a hook for his right hand I believe. And anyway the girl's like freaking out, like let's get outta here, let's get outta here, and so finally he's like OK we'll leave and they got to wherever they were going and cause he was a gentleman he got out of the car and he like walked around the car to go open the door for her to let her out and when he went to her door to go let her out, all he saw on the door handle was a bloody hook. And that's all I remember. That's it.

Despite the loss of various important features, the bounded phrases that define the underlying narrative structure are still in order, and the general notion of threat and its resolution have been maintained (see fig. 3 below). The last two links in the chain were, in contrast, reluctant storytellers at best, and the story quickly lost any coherence once it reached these students:

C₁₁: This was in Massachusetts somewhere. And um so there were these college kids and I guess they wanted to just go away for the weekend, so they were going to this lake kind of far away, they'd been driving all day and so while they were driving they, uh, someone on the radio came on, and they were like warning everyone, because this like insane murderer had escaped from like the mental hospital and it was just in that area where they were, so they uh, they decided that uh they should probably like turn around and go home. And when they came home there was a hook on the door of the car. And that's it.

C₁₂: OK. Well there's some college students who are from Massachusetts and they decided to go to a lake, and then they went to the lake. And they left their car I guess to go onto the lake for some reason. When they came back there was a hook on the car and they were very scared.

Some simple metrics can help reveal the rapid changes in length and word choice. Below is a table tabulating the most frequent verbs and nouns in the story, as well as a tabulation of total word tokens and word types for each narrative (fig. 2):

Perhaps most striking is the rapid decrease in both word tokens and overall vocabulary. While the verb list shows little of interest, except confirming the past tense nature of the narration, the noun list does highlight some interesting phenomena. Most obvious is the rapid disappearance of place-name referents—the story was deliberately set in a landscape that few of the students had experience with and, consequently, the very specific and unusual place referents disappeared almost immediately, to be replaced initially by more generalized place referents, and later to drop out altogether. Place referents in the network story were far more persistent (see below), both because they were well-known to the students, and because the students heard and performed the story numerous times, helping to fix the story-place link in their memories.

Another fairly straight forward comparison method between these stories is the isolation of what Labov and Waletzky (1967) term "bounded phrases." Below is a small table that attempts to align the bounded phrases from each of the four variants in the order in which they appear in the stories. The last variant is particularly challenging, as the narrative bears little resemblance to the original narrative (fig. 3):

Although the underlying structure—and the order of bounded phrases—remains constant across the variants, the degree of detail between each telling drops considerably. Of particular note is the complete disappearance of the complicating action of the story (the escape of the insane man, identifiable by his hook) in C₁₂ (Labov and Waletzky 1967). Clearly, for the chain story, the lack of any reinforcement mechanisms, coupled with the lack of cultural relevance for the students, doomed the story to significant instability and, ultimately, a degree of incoherence that would clearly prevent further telling of this story.

Anderson (1951), in his classic storytelling experiments, recognized the importance of network transmission—a situation that more closely resembles an active tradition community in which stories are told and retold by numerous storytellers.⁵ As an extension of my class's chain experiment, we decided to develop a "seeded network" experiment in the following quarter, where I provided a story to two members of the network and then enforced a network transmission over the course of nine weeks with eight participants (fig. 4).

Students were asked to retell the story each time as best as they could remember it. Transcripts of the first telling and the last (fourth) telling of the story by a single individual provide an interesting comparison:

(A₁) So this guy moves out to Los Angeles to UCLA from the East Coast and he's never been out here before and he doesn't have a car and he realizes that he'll need a car to get around because it's Los Angeles and public transportation just really stinks here. So he's looking through the paper and sees this listing that says brand new Porsche used for sale, for sale for 50 bucks and he knows that this is a typo but (…) "Ah, what the hell, I'll call up anyways." So he calls the number and "Oh, you're calling about the car, fantastic, um, come on out for a test drive," and he says, "Oh, sure." So, you know, this is a once in a life time opportunity to drive a Porsche so he goes out to the address, this big estate in Beverly Hills, right 90210, right in the driveway is this glowing, brand new, red Porsche. And he can't believe how beautiful the car is and wishes he could have it, but he knows it's out of his price range. He figures, you know, I'd love to drive it anyway so he knocks on the door. "Oh great, you're here about the car, let me show it to you, let me show it to you." He walks around, she shows him all the features, the interior design, the beautiful leather interior. "Well, do you want to take it for a spin?" So they hop in, he's driving around, driving for hours up and down the Coast Highway, and he just loves it. He knows the car is getting low on gas so he returns. He says, "Well, look, you know, it looks like I've used up all your gas, so I'll reimburse you for the gas but I, I really can't afford the car." And she says, "Can't afford it? But it's only 50 bucks! What do you mean you can't afford it?" "Well, gosh, I, I can afford that! Why are you selling it for so cheap?" And she said, "Oh, well, my husband left me for his secretary and I just got a wire from Barbados saying, 'Please sell the car and send me the money'."

(A₄) This friend of a friend of mine moved out from the East Coast to discover that here in LA it's impossible to get around without a car. But he's kind of a starving student so he doesn't have a lot of money to afford to buy a car. So he's looking through the paper and he sees this listing for this brand new Porsche for 50 bucks and he's like, "Ah crap, it's definitely a typo but, you know, I gotta go, I gotta go test it out, you know, I gotta go." So he hops on the bus, heads out to the address and it turns out to be this big mansion in Beverly Hills. So, so he knocks on the door, he notices that there's this beautiful, red Porsche out front, he says, "Oh, god, I would love to have that car, but at least I'll get to sit in it, I'll get to see it." A woman comes to the door, "Oh, hey, great, so you're here about the car, fantastic! Let me show it to you." So they walk out, looks at it, opens the door, smells the leather and gets in. "Hey, you want to take it for a spin?" Says, "Oh yeah, sure, that'd be great." So, you know, they drive around Beverly Hills for a while and they take it on PCH, driving around for hours and he notices he's kinda getting low on gas so he, "Ah, well, you know, it's been a long time and we're low on gas," and he starts to bring the car back. So they pull in and he's like, "You know, I, I'm sorry, I'll give you, I'll pay you back for the gas, but you know I just can't afford this car." And she's, "You can't afford it? What do you mean? It's only 50 bucks!" And he's like, "Holy crap! Oh, ok, here I got 50 bucks." And so she hands him the pink slip, and he's thinking, "OK, I gotta know now that I can ask her." "So why are you selling the car for 50 bucks?" She says, "Oh well, you, my husband ran off with his secretary for Barbados and he sent me an email the other day saying sell the car and send him the proceeds."

These two story variants reflect many of the aspects of crystallization that I identified in an earlier study of paramedic narrative (Tangherlini 2003).

Similarly, the first telling in the network and the last telling in the network by different storytellers reveal equally interesting aspects of stability and variation:

(B₁) So the story is, ok, a guy moves to LA and sees an ad for a Porsche Carrera and it's for 50 bucks. But he thinks that the K got left off of the ad and he figures what the heck I'm test driving cars anyway and I have time to kill. So he call this 310 area code and a woman answers and says to come and test drive. The address is in Beverly Hills 90210 zip code and, uh, he takes the bus to Wilshire and Rexford and walks to this big, old house. It's a huge mansion—it has a circular drive way and there's a Porsche Carrera, cherry, it's bright cherry red and (…) and he takes the car. Test drives it with the woman. Goes on Mulholland, she doesn't seem to mind his driving very fast and he drives for… he drives to PCH…has gone for an hour and then they're gone for two hours and then when it gets to the point where he's low on gas, then he goes back to Beverly Hills. He says at this point, "I wish I could afford this car," and the woman is astonished and says, "What do you mean, you can't afford 50 bucks?" and so, not to scotch the deal, without saying anything anymore he gives her 50 bucks and she gives him the pink slip. But he can't resist asking why she is selling the car for 50 bucks. And she says her husband ran away to the Caribbean with his secretary and told her to sell the car and send him the money.

(G₄) So this is the story of a student who was from the East Coast and he comes to the West Coast and he's been told that, well, things are not like in New York, you cannot just hop into a bus or a metro here, unfortunately, you need a car. So he was looking in the newspaper for cars and he saw this ad for a Porsche for 50 bucks—he thought it was kind of odd, a Porsche at that price, but he said, "What the heck, why not give it a try?" and he called the person and see what's up. So he takes his bus, sorry, he takes his bike up to the Beverly Hills and the address that corresponds to it is an absolutely gorgeous house. Very, very nice house, and in front of it there's this beautiful, red Porsche, the one that everybody dreams of. And then he goes up and there's this very nice lady who welcomes him and she offers him to take a spin. So of course they go and they go for a ride and it's the smoothest ride he guess he can have, it's just marvelous, he loves that and they drive for half an hour up to PCH and he's just in heaven, it's beautiful inside and then the gas is a little low, so he goes back towards the lady's address. And, um, at the time he has to take a decision to buy the car he says, "Well, I'm very sorry. I don't think I'll be able to pay for this car, I can't afford it." But the woman looks at him surprised, saying, "Are you mad? I mean, this is only 50 bucks! If you can't pay 50 bucks for a car…" and he doesn't ask more, just hands over a $50 bill. She gives him the pink slip and later on, a few seconds later, he says, "Thank you very much for this but I'm curious to know why, uh, this Porsche is so low in price." So the woman says, "Well my husband left last week with his secretary for Bermuda or those islands and he just sent a message, email or phone, saying, 'Sell the Porsche and send me the money'."

Not surprisingly, the network, with its multiple performances and multiple opportunities to hear the story—and thus reinforce the memory of the story—leads to a far greater degree of stability in the story. Again, the same short analytical illustrations as used in the chain story reveal a remarkable degree of stability across the network, a stability that aligns well with my study of crystallization in paramedic personal experience narrative (Tangherlini 2003). The word frequency table reveals not only a consistency within narrative repertoire (here narrator A), but across the network as well, both for word choice, total vocabulary, and length of narrative (measured in word tokens) (fig. 5):

The stability within repertoire according to these criteria is particularly striking and obtains for all of the storytellers in the study.

A second comparison of bounded phrases is equally revealing. The alignment of phrases is much easier in this case than in the chain transmission. This stability can most likely be attributed to the reinforcement mechanisms of the network transmission of the story (fig. 6):

Although the word choice does vary (as noted in fig. 5 above), the overall structure of the story varies very little. Given this stability, with more retellings, it is likely that these storytellers would be able to pick up their stories at various points in the telling, without the need for a "running start." The slight changes in formulation both within and across repertoire suggest that the story is readily available to introspection during its retelling. Repeated exposure to the story of course leads to subtle changes in each telling, undermining the idea from the Standard Theory that, once consolidated, a memory is unlikely to be changed; rather, here the environment of repeated telling and shifting information changes the underlying memory itself at the same time as it solidifies the pathways to that memory.

A clear limitation of these experimental observations is that they do not allow for an assessment of the ability of storytellers in the network to begin their story from points other than the beginning. Rather, the experiment simply confirms the widely-held notion that networks are better for maintaining stability of memory than chains. Although this begins to address the question of how people learn, store, remember and perform traditional expression, the premises of the Standard Theory could possibly account for these observations. Yet MTT provides a much better model for understanding the shifts within repertoire and across a network as hinted at in the above experimental data. It also does a much better job of accounting for change over time. The next step is to devise a network that allows for interruptions, yet requires the storytellers to continue their story with, or without, a running start. In such an experiment, one would probably need to split the storytellers into two groups. Group A would be allowed a running start and Group B would be asked to tell from where they believed they left off. If the discussion above concerning MTT and the "craftsmen of tradition" is correct, the experimental group would have to include several such "craftsmen"; as such, it might be difficult to find an adequate tradition group.⁶

My brief experience as a neurosurgeon's consultant obviously led me down what I hope will be a fruitful path in understanding both the physiology and the neurobiology of tradition. Advances in neurobiology—and a move away from viewing the brain as a "black box"—now allow folklorists to consider the physiological structures and the biological processes that make traditional expression not only possible but guarantee that they will continue to be performed. Tradition, it turns out, is conditioned by processes of learning, memory consolidation, and memory recall. The recognition that every brain is different—both because of genetics and because of the dynamic impact of physical and social environment on the brain—fits well into a view of folklore as emerging from the dialectic tension between the individual and tradition. Understanding the neurobiology of learning and memory further clarifies the basis for stability and variation in tradition. MTT provides for a dynamism in long-term, declarative memory missing from the earlier Standard Theory of memory consolidation. Similarly, an understanding of the relationship between declarative and procedural memory explains the relationship between performance and text; to borrow from Alan Dundes (1964), there clearly is a neurobiological basis for texture, text and context. While it may be too early—or perhaps misleading or even wrong—to posit the notion of "tradition memory," it seems quite clear that recent advances in understanding long-term declarative and nondeclarative memory also can help clarify intriguing phenomena with regard to the performance of traditional expressions that appear both during fieldwork and in the archive.