April 24, 2016

Remembering Life Stories (5): Biographical Expertise

our ability to “chunk” elaborate sequences into single units of mnemonic content is dramatically enhanced by extreme familiarity with diverse variations of common biographical patterns

Remembering is constructive, but in more ways than one. Beyond the famous “schema theory” of Frederick Bartlett (in which “gaps” are filled in due to subconscious prejudice) we’ve also discussed “memory for the time of events” according to William Friedman (in which “dots” are connected via contextual implications). Now, in my last post, I’ve introduced what might be considered a third manner of mnemonic reconstruction, in which a familiar serial pattern is remembered constructively, piece by piece, and yet recognized as a whole. Even if this typology is not yet as crisp as I’d like, one thing does seem clear. Common patterns of Biographical Temporality can provide ready made sequences which can help structure a reader’s mnemonic fabula of a Life Story. A familiar serial pattern is mnemonically compressed into one “chunk”, one single “unit” of content. Because of this cognitive process, any reader familiar with Biographical Patterns of Temporality is naturally prone to find greater coherence in Biographical Storylines.

This contention demands further scrutiny. One challenging question we’ve not yet asked has to do with complexity. For a given reader, how many Life Story patterns can be chunked as units (and thus, efficiently reconstructable)? How many of these patterns can individually be recognized as separate and distinct whole units? If “unit-izing” a single pattern requires strong familiarity, how can a general readership be “familiar” with a broad diversity of various patterns?

In part one of this series, I said we should never suppose all readers’ minds generically or uniformly employ the same general pattern of biographical content, such as Burridge’s stages of composed ancient ‘bioi’. In parts two and three, I suggested that probability and statistical awareness can support any particular pattern a given mind might so happen to employ. In today’s post my goal is to mediate between these two points. How can a multitude of distinct biographical patterns be available, each one being as unique as some particular life story might require, and yet also being general enough that a given biography can be remembered coherently by a broad audience of individual readers? How many biographical patterns can the mind possibly recognize as unified life stories?

The answer depends not on the strength of an individual mind, but its orientation. The volume and variety of familiar sequences a mind can learn (i.e., “chunk” into long-term memory) depends on the studied interests of that particular mind. A depth of interest and years of paying attention to one subject - the redundant encoding of repeated observation over countless iterations - produces what psychologists call “expertise”. In short, we differentiate biographical patterns in memory because we have all become expert observers of basic human behavior.

As Carlyle first pointed out, people are obsessed with people. There is no other subject to which we pay such close and constant attention. Thus, generally by adulthood, biographical patterns become our area of shared expertise.


Psychological discussions of expertise usually began with the famous studies of chess masters, begun by Adriaan de Groot in the 1940’s and followed by William Chase & Herbert Simon in the 1970’s. Surprisingly, instead of higher IQ’s or superior mnemonic ability in general, de Groot found that what distinguished chess masters from novice players was their studied familiarity with the patterns of the game. For example, when asked to memorize a chess board with some chess pieces laid out in random positions, a grand master and a beginner would fare equally well. However, when asked to reconstruct board layouts from actual game play, the expert players were far superior in both speed and accuracy. It was Chase & Simon (drawing on G.A. Miller) who first realized that experienced players were reconstructing the board in chunks, by because they remembered the placement of pieces by groupings, rather than individually. By contrast, without a sense of the structural contexts involved in situations of actual gameplay, an inexperienced player can only attempt to memorize pieces one at a time. To a novice, every placement seems random. To the expert, complex patterns are not only familiar, but integral.

What is most significant here are the contrasts. First, it takes familiarity in a domain to observe patterns amidst the arbitrary. Chess is definitely arbitrary. The total possible layouts during actual chess games is a number that begins with “1” and ends with dozens of zeroes. Within those possibilities, however, the structural contours and basic rules of gameplay guarantee statistically that some layouts are going to be far more common than others. That means frequent patterns are there to be found, amidst all the arbitrariness, but those patterns cannot be noticed until players develop expert level familiarity. Second, the phenomenon of expertise is domain-specific. Chess masters are not exceptional at remembering in general. They only possess superior chunking and reconstructive abilities in their area of expertise. When psychologists discuss “expertise”, this domain-specificity is always implied.

All this was well summarized in the Oxford Handbook of Memory by Robert Lockhart. “It indicates that memory superiority is limited to the area of expertise and that the impact of expertise on memory lies in the expert’s ability to extract meaningful patterns from stimuli that to the novice are effectively random.” In short, we most easily recognize things we already know.

Aside from Chess, the cognitive benefits of specific expertise have been documented in other areas such as music, science, sports, and medicine, most famously by the “expert expert” Anders Ericsson. While Ericsson’s work has focused on task oriented performance, rather than information oriented recognition, and although he prefers to write in terms of knowledge and learning rather than memory and chunking, the core principles stand in alignment. For instance, in one 1999 article, Ericsson notes that experts demonstrate superior quality in their mental representations, and these “acquired representations appear to be essential for experts’ ability to monitor and evaluate [and improve] their own performance” (MITECS p.299). Surely this process of acquiring and leveraging “mental representations” is equivalent to cognitive chunking and its applications in constructive remembering, because expert level performance relies on first having built up an extensive knowledge of the domain. This is self-evident, for instance, as much in the number of chords and progressions learned by musicians before they surpass Ericsson’s “ten thousand hour” benchmark as it is for the detailed variations learned by experienced dancers which enable them to pick up new choreography quickly. Indeed, the same holds for learning a second language; those who pay rigorous attention to mastering patterns of grammar and syntax will acquire the ability to apply that knowledge (to employ their memory of those serial chunks) in speaking and reading fluently at advanced levels.

In addition to Chess, and other areas discussed by researchers like Ericsson, I submit that expertise applies also to this domain of Biography, and at this point we can once again start to get practical.

The reason Biographical Expertise is such an advantage for Remembering Life Stories is simply that readers have already done the hard work of learning various sequences before they ever begin reading. That is, such readers have previously spent their whole lives cognitively chunking a multitude of acutely differentiated domain-specific serial patterns which represent commonly observed sequences in human behavior, and this includes all the “time patterns” Friedman accounted for as “general knowledge”. For example, all the illustrations of temporal patterns I offered in post #3 are most likely embedded right now in the remembering minds of whoever buys next month’s most popular biography.

It’s worth noting here, from a hermeneutical point of view, that whether the arbitrary sequence of a particular Biography can be said to contain non-random subsequences is dependent upon prior audience knowledge, which means the “Unity” of a Life Story depends on information not in the text. Traditionally, such a prospect has been hermeneutically dauting, and I cannot help suggesting this may be one factor that helps explain why critical theorists of historical narrative have traditionally quarrantined Biography as a genre.

What’s most challenging about all this, at the moment, is attempting to grasp just how immense and how vastly differentiatied this collection of biographical patterns must be. How can it possibly be true that so many distinct sequences can become “familiar”, even to experts?

Consider the label “education & career”. Whether this counts as one concept or as two put together, the number of actual sequences entailed by “education and career” can be endlessly diversified. On the education side, there is elementary school, middle school, high school. Or perhaps some particular mind focused more specifically on each individual grade (K, 1st, 2nd, etc). There are also variations less common but not uncommon, such as dropping out and returning to school, or dropping out and passing the general equivalency test, or being homeschooled, or being educated on various army bases during a parent’s multiple deployments, or just taking a year off due to illness or circumstance. And so on and so on, ad infinitum. On the career side, again, variations abound. Many careers begin with graduation, but some careers begin before commencement. Entrepreneurs and professional athletes are only two of the more high profile examples. So there is sometimes “startup year, graduation, second year…” Then there are internships, a distinctly different pattern of unpaid labor in between graduation and entry-level career opportunities, which becomes yet another unique sequence when a summer internship takes place one or two years prior to graduation. There is also an endless variety of patterns within individual careers. For some blue color job tracks, there are strict thresholds for “rookie, journeyman, veteran”. In other fields, sadly, it can often be typical for an entire career sequence to be defined by a series of entry level jobs that never result in advancement. You, dear reader, may not have known such workers intimately, but you’ve interacted with literally thousands of cashiers, and you’ve easily had hundreds of chances to observe similar low-level workers interact with their managers.

This incredible diversity of biographical sequences seems to defy formal classification as a collection of patterns. From a wide lens perspective, it looks more like chaos. But then, consider how chaotic the English language must have seemed before Samuel Johnson. How many domain-specific patterns can one expert learn to recognize? Well, how many words do you know in the dictionary? If you can easily see how the letters “a-m-a-z-i-n-g-l-y” have been chunked in your own memory, then multiply that experience by the hundreds of thousands. The Oxford English Dictionary contains over 171,000 active words, 47,000 archaic words, and nearly ten thousand more with derivative sub-entries. If we also count plurals and forms in the past tense as distinct variations (which of course technically they are) and an avid reader might conceivably chunk 300,000 letter sequences in the course of a lifetime, if not perhaps even more. Just consider the variations “farm, farmed, farms, farmers, farmer’s, farmers’, farming, farmland” and ask yourself if the diversity of biographical sequences is comparatively worse.

How many variations of serial patterns can the mind actually establish as familiar? How many biographical patterns can we realistically remember? In theory, given enough patterned content and domain-specific expertise, we are each equally capable of recognizing (and “unit-izing”) as many serial patterns of familiar human behavior as the numbers of words we can spell. English writers long before Samuel Johnson had formalized the spelling of several thousand words. Over 30,000 distinct “chunks” of unitized content are displayed in Shakespeare’s writing alone, all of which - without a dictionary - he had to spell by heart. Critically, it’s beside the point to wonder if Shakespeare’s audience could do more than sound out a few basic spellings. The Bard himself had long passed through the requisite stages of acquiring language. His expertise was established and his chunked vocabulary was massive. In theory, we can all develop a similar volume of chunks in our repertoire. We can all develop increased powers of remembering, even of multiple groupings and sequences as whole units, provided that such patterned content belongs to a personal area of expertise.

More to the point, we have already developed (to our own personalized degree) our own relative expertise in observing patterns of biographical development.

Going back to the “farm, farmed, farms, farmers…” example, it might be considered helpful to realize that a large portion of the new patterns we typically chunk are no more than slight variations of patterns we’ve previously chunked. Along these lines, some psychologists expand chunking theory into something called Template Theory. The basic idea behind template theory is that a familiar sequence (“On top of old smokey, all covered in snow”) can facilitate the quick learning of a slightly modified sequence (“On top of spaghetti, all covered in cheese”). Test yourself quickly, please, by attempting to memorize these four 17 word strings:

(1) Beulah asked Millicent to expunge the problematic automotive accounts and to validate Jerilyn’s mediocre draft proposal. (2) Take Johnson Highway until Henderson Trail. Take the next two left turns and you’re five miles away. (3) I came, I saw, I conquered, and I’d appreciate a little courtesy and respect for my trouble. (4) Ask not what your mother can do for you. Ask what you can do for your mother.

Notice how each verbal sequence is more patterned than the one before it. In the third string, you had the first six words memorized as soon as you’d read them, and the remaining portion was made up of common phrases that seem familiar together. However, the fourth string offered you an entire “Template” with one slight change. Where JFK had said “country”, we insert “mother”. I dare say you should have memorized that last one somewhat instantly. To round off this digression, template theory is one way of accounting for the more acute differentiations involved in learning a multitude of differentiated patterns.

There’s no question, dear reader, that your own expertise for such things has grown continuously during each decade of your own life experience. You’ve had to learn literally millions of patterns, but you’ve also had approximately two hundred million seconds of waking attention per decade to devote to such learning. You’ve watched, noticed, and thought about other people’s lives, consciously and subconsciously, probably more than you’ve done anything else. Thus, to some degree or another, you are all certifiable experts, by now, in observing biographical sequences.
Your biographical expertise makes it easy to diversify your own knowledge of biographical patterns, and on some level you have already done this for lots and lots and lots and lots of these patterns. It may seem incredible to think we have learned countless variations in biographical sequence, since we tend to think of each one needing to be learned individually before it can be individually familiar, but that’s not how the process actually works. In practice, a lot of patterns are similar.

You learned to read and to recognize spellings of words as the result of a complicated process (about which psychologists don’t entirely agree) but the point at which you began to master spelling as a discipline was after you had thousands of opportunities to recognize common patterns in word structure. Think back. Nobody sat down and explained to you, then, that our most common vowel is E, or that T is most often followed by H, or that N is most often followed by a T or a D. They might once have mentioned that Q is invariably followed by U, but written English displays hundreds of relevant digraphs and trigraphs. What actually happened was consciously imperceptible, but on some level the frequent pairings and triplets were a benefit in your learning process. The common digraphs (TH, QU, NT, ND, etc) and trigraphs (THA, ENT, ION, TIO, etc) helped you learn to read and spell with proficiency.

So it has been, all your life, with biographical sequences.

General readers already possess expertise in observing human experience, which allows readers’ minds to recognize various longitudinal patterns of behavior in narrative literature. Because our minds have long ago chunked a collection of similar patterns as sequentially ordered groupings of mnemonic content, our constructive remembering of (the chronological fabula of) narrated life stories is significantly advantaged. Even when life stories eschew narrativized emplotment, we remain capable of receiving such stories coherently despite their lengthy, elaborate, and arbitrary storylines. We can do this because biographical content is essentially self-structuring, and because the vast bulk of biographical structures are themselves self-contained (pre-structured, organized, chunked, “unit-ized”).
In short, our ability to constructively remember lengthy, elaborate sequences is greatly advantaged when working with biographical content.

As it so happens, Thomas Carlyle had inadvertently tapped into more than he knew.


It must be emphasized once again that chunking serial patterns (or “templates”) is a profoundly different experience than utilizing frames or scripts or schemata. Frederic Bartlett’s signature experiment was based on his careful and deliberate selection of a bizarre little story, “The War of the Ghosts”. Shrewdly counting on the forgetability of that weirdness is the primary reason why Bartlett was able to observe such a wealth of contrived rationalizations. To be sure, his conclusions (when observed rightly) remain unassailable. Readers’ minds tend to replace unfamiliar details with other material that aligns better with personal expectations or conforms more with general & conventional experiences. We tend to remember creatively when story content is lost, and we often fail to preserve data for which we have no prior context.

By the same token, however, our ability to reproduce stories mnemonically should increase when material happens to fit well with familiar contexts. Indeed, Bartlett’s own summary of conclusions from that signature study includes statements like “Detail is outstanding when it fits in with a subject’s pre-formed interests and tendencies.” and “the reduction of material to a form that can be readily and ‘satisfyingly’ dealt with is very prominent. [This process] gives the whole dealt with that specific ground, frame, or setting, without which it will not be persistently remembered” (Remembering, p.92). In other words, one could technically argue that chunking serial patterns is very much like schema theory in their core concepts and functions. However, I must insist we continue to emphasize their distinctions, and for two particular reasons.

First, the connotations of scholarly discourse about schema theory have overwhelmingly focused on what happens when information is lost from memory, rather than what happens when information is preserved (Ost & Costall, 2002). People think of schema theory as imposing familiar contexts that don’t necessarily apply. That’s not remotely what happens with chunking and specific domain expertise. Second, schema theory typically speaks of stereotypes and generalities. Schemas are gist memory contexts that may or may not include details. Likewise, “cognitive scripts” are far more generalized as sequential experiences than the types of biographical sequences our minds tend to chunk as serial patterns. Scripts and schemata do not offer the kind of variation we find with expert chunking (with or without “template theory”).

In sum, Bartlett was right to teach us that unfamiliar content is usually replaced, but one must therefore also suppose familiar content is often retained. That is, especially if by “familiar” we mean the levels of expertise illustrated in this post, above.

The first point to recap is that sequential “chunks” of story content can be mnemonically unified. A familiar serial pattern enables multiple pieces of information to combine as one single unit of memory. In other words, defying Aristotle’s opinion, we may justifiably declare that biographical patterns possess a degree of Unity. Although less comprehensive than the unity of a classical plot, Life Stories can and do enter audience memory with a sense of wholeness. This is a cognitive phenomenon, occurring when readers recognize a familiar serial pattern in material that reflects biographical temporality, enabling that pattern to be more easily remembered altogether - a sequential Unit which incorporates many parts in one particular whole.

The most critical problem this post has tried to address is that all this requires decades of learning. This kind of remembering requires not common knowledge, but focused expertise, which is where Thomas Carlyle’s intuition comes back into play.  We have all been obsessively focused on learning about other humans’ experiences.

In principle, a biographical storyline can be individually arbitrary and yet resemble a temporal pattern that is well known to some readers. As often as this happens, the “arbitrary” storyline can be remembered coherently - at least by those particular readers. We must note once again how none of this reflects any sense of causality. As we’ve seen in posts #2 and #3, the reader’s familiarity can be purely statistical. If some generic “event sequence” is easily recognizable as a frequent set of consecutive outcomes, then any storyline evoking that sequence is more efficiently rememberable. That is to say, patterns increase coherence.

Because our Biographical Expertise enables us to recognize many diverse patterns as individual units, we have an increased ability to maintain coherence when Remembering Life Stories.


In my next post, we’ll reassess biographical patterns from a statistical point of view. If we redefine common patterns as collections of frequencies, we can re-examine the coherence of biographical sequences in more informational terms, as statistical regularities.

It has been said that Plots (in hindsight) are always predictable, Biographies are arbitrary, and Chronicles are random. But my next post will declare these are not three distinct categories. They are portions of a continuum. The truth is, Plots, Biographies, and Chronicles all present various degrees of coherence, but that coherence is not literary. It is mnemonic. It is, more precisely, mnemonically reconstructive. Whether aided by causality or probability, some discourses stories are more coherent (i.e., more reconstructable) than others.

Altogether, this suggests that mnemonic encoding of stories (storylines) can be hypothetically measurable, according to (what I’m going to call) Narrative Redundancy.



Addendum: on Biographical Patterns and Historical Theory

We might here briefly revisit the conflict between Louis Mink and David Carr. With Mink, I say the non-fiction stories we reconstruct in our minds are not accurate accounts of the past, but against Mink, I say that life experiences are cognitively compressed into a storyline by the natural process of autobiographical memory. Against Carr, I say that human experience itself does not posses an inherent narrative structure, but with sympathy for Carr (whose arguments could be modified without too much trouble into cognitive terms) I say that our minds naturally create “chunks” of meaning and structure from whatever information gets past the “filtering and boosting” function of our “attentional system” (Daniel Bor, The Ravenous Brain, p.126-7).

In other words, when I say that “our minds have long ago chunked patterns as sequentially ordered groupings of mnemonic content”, I do not mean to suggest that these patterns are necessarily accurate models of any human experience in particular, or even in general. What I am saying is that these patterns are accurate representations of the way that similar samples of lived experience (first, second or third hand experience) have been processed over time by our cognitive faculties. The familiar sequence of “high school, college, entry-level career opening” is not strictly an accurate account of anyone’s life. Technically, that sequence admits an astounding level of distortion and commits countless sins of omission. We cannot stress strongly enough that Louis O. Mink was absolutely correct on this point.

On the other hand, we do not contradict Mink if we say the familiar sequence of “high school, college, entry-level career opening” is an accurate account of a common distortion, a mnemonic compression which occurs frequently in human remembering. In short, while Carr remains absolutely mistaken to insist “human experience” has an inherent narrative structure, he would not have been incorrect to insist that life often seems that way in retrospect.

Car was incorrect because he tried to identify stories with reality. It might have helped him to realize how significantly his own view of reality had been distorted by personal remembering.

Some biographical sequences become familiar precisely because the activities they distort and the experiences they compress display a common pattern in long term human behavior, which is thus a common source of our resulting memories. These sequences thus do represent the past “accurately”, albeit within some degree of distortion. Human activity does not have an inherent narrative structure, but our cognitive nature - the consciousness which must observe things in linear order and the encoded content that implies temporality in constructive remembering - is consistently at work making experience seem like a narrative in retrospect.

Lived experience is indeed very much unlike a story…
Except as it seems afterwards, on the inside of our minds.


No comments:

Recent Posts
Recent Posts Widget
"If I have ever made any valuable discoveries, it has been owing more to patient observation than to any other reason."

-- Isaac Newton