The “Umbrella Man”, Occam’s Razor, and the Archaeological Noise of Personal Practice

Abstract

This essay considers the ramifications of “The Umbrella Man” - a brief mystery that was of interest to historians and theorists examining the assassination of John F. Kennedy in 1963 - on the practice of modern archaeology and some of the basic tenets upon which it relies. The paper focuses on certain materials, analytical methods, and research themes associated with the study of nineteenth-century sites in the Midwest, contextualized to varying degrees by the presence of a written record left behind by the makers and users of the material world as it existed in the region during the 1800s.

The “Umbrella Man” refers to a man who can be seen in the Zapruder film taken in Dallas in 1963, where Kennedy was assassinated. This formally dressed figure can be seen holding an open umbrella on a bright, sunny day – the only person in the entire crowd with an umbrella (Fig. 1). Further, he was standing at the very point on the street where shots began to ring out on the President’s motorcade.

Fig. 1

The Umbrella Man in Dallas, 1963

Historian Josiah Thompson discovered (and named) the Umbrella Man. In doing so he noted something that a) stood out as significant and b) pointed out a circumstance that suggested - in its improbability - a certain apparent (and in this case, malicious) conclusion (Thompson 1967). Thompson took care not to draw firm conclusions himself, but simply pointed to the anomaly and stood back to watch various theorists have a virtual field day with conspiratorial interpretations (Morris 2011). One such theorist actually created a schematic drawing of a proposed ballistic umbrella, which would have fired upon the motorcade as it passed.

John Updike (1967) put this event into an interesting perspective:

We wonder whether a genuine mystery is being concealed here or whether any similar scrutiny of a minute section of time and space would yield similar strangenesses — gaps, inconsistencies, warps, and bubbles in the surface of circumstance. Perhaps, as with the elements of matter, investigation passes a threshold of common sense and enters a sub-atomic realm where laws are mocked, where persons have the life-span of beta particles and the transparency of neutrinos, and where a rough kind of averaging out must substitute for the absolute truth. The truth about those seconds in Dallas is especially elusive; the search for it seems to demonstrate how perilously empiricism verges on magic.

Ultimately, and improbably, the mystery was solved. The man with the umbrella was Louis Steven Witt. He explained that he donned the umbrella simply to heckle Kennedy, whose father had been a supporter of the Nazi-appeasing British Prime Minister Neville Chamberlain who often sported a black umbrella as a fashion accessory. His obscure protest, and his presence at the point of assassination, was a misleading (and bizarre) coincidence.

But there is still an important take away here. The man standing in the sun with an umbrella, next to the motorcade and at the very place where the shots began to be fired, suggests a valid assumption that his anomalous dress and his position was somehow related to an anomalous event that happened at the same place and time. It was a reasonably sound assumption – and a simple one – to suggest a relationship between the apparently inexplicable Umbrella Man and the sudden assassination.

The simple, logical connection between the gentleman with the umbrella and the assassination, however surreal, could be regarded as an appropriate engagement of “Occam’s Razor.” This principle, often paraphrased as “the simplest solution is most likely the correct one,” suggests that when presented with competing hypotheses, the most sound solution is that with the fewest assumptions.

The concept of Occam’s Razor is often used in the classroom or laboratory as an important reminder to students or inexperienced analysts to consider first the conservative interpretation of the function and meaning of archaeological remains. The elaborate or unique interpretation for the use or meaning of an object, however ingenious, must be weighed against the probability of a less elaborate, more prosaic interpretation. Generally, this perspective is a safe and reliable one. As archaeologists, we are charged with a number of tasks in the lab, all of which are inherently dependent on basic logical interpretations and the implementation of Occam’s Razor when confronted with a minor conundrum.

Once the broken things are cleaned, labeled, and bagged, we generally focus first on the creation of basic chronologies and histories – of the occupation of a particular place on the landscape, and more broadly, of technologies, practices, or of societies themselves. All of these are based (presumably) on the accurate interpretation of the design, function, and implementation of the lost or discarded items that we know as artifacts.

The design, function, and use of an object can be deduced. This object is made of glass. It is a bottle. It held something. Based on certain attributes of that bottle, it held this substance. And that substance was used by its consumers for this reason. And that use represents a range of pragmatic needs and abstractly motivated choices.

Building on these interpretations, we can reach further into the lives of the those who left these objects behind, attempting to interpret patterns that reflect personal or cultural practices or traditions. As we continue to draw broader conclusions about past peoples or events, we rely more and more on the soundness of the foundation created by the interpretations and assumptions surrounding dozens, or hundreds, or thousands of individual artifacts. It is here, then, where we trust the general reliability of Occam’s Razor – those simple, logical assumptions about function and meaning that can burden the weight of additional assumptions on which they will stacked.

Lurking Umbrella Men

However, even within the familiar realm of mass-produced, still-familiar objects of the nineteenth century, there may lurk “Umbrella Men” waiting to render Occam’s Razor incorrect. Often, this reflects user appropriation or the renaming of objects designed for a different function – an act that (unfortunately) is generally invisible archaeologically.

For instance, during the mid-nineteenth century, a number of glass factories in Louisville, Kentucky (and elsewhere) produced distinctive, fiddle-shaped bottles known as “scroll flasks” that were intended for the sale of whiskey. These were sold empty to local retailers who then filled them with a variety of whiskeys and spirits, often attaching a small paper label to identify the contents. The intended function of these objects is well documented, and they are ubiquitous in archaeological samples from across the American Midwest (e.g., McKearin and Wilson 1978).

Recently, a non-archaeological example of a scroll flask was found in an antique store in Illinois. Attached to the exterior of the bottle is a small paper label that clearly dates to the mid-nineteenth century. That label reads “turpentine.” Had this bottle been buried in the ground, that label would have disintegrated quickly. But this non-archaeological example demonstrates that the retailers who purchased these empty, distinctive bottles, were free to use them as they wished. The meaning of this whiskey flask was altered by the merchant and the customer who purchased it. A whiskey flask became a turpentine bottle. Thus, any conclusions drawn from its morphology alone (aside perhaps from its age or place of manufacture) would be incorrect in an archaeological context. The flask becomes a kind of Umbrella Man. The most logical conclusion is false, and the actual conclusion is nearly impossible to predict.

The written record also provides instances where users modified the meanings of mass-produced goods. A settler of the 1830s Illinois frontier complained that she had broken three Staffordshire tea saucers while using them as grease lamps (Burlend and Burlend 1968). A British traveler in the Southeast, when requesting a chamber pot for his room, was to his horror handed a kitchen kettle (McWhiney 1988).

One of the most abstract examples of an archaeological conundrum akin to the presence of the man with the umbrella in Dallas comes from the author’s own personal experience. During the mid-1990s, I was excavating the remains of a pit cellar associated with an early nineteenth-century tavern in the small village of Sangamo Town in central Illinois. A young Abraham Lincoln had spent time there during the summer of 1831. It was abandoned by 1835 (Mazrim 2007).

On the floor of the tavern, clearly discarded while the building was still standing and before the clay walls began to slump into the old facility, was a pearlware saucer. Or what appeared to be a saucer. It was resting face down on the clay floor surrounded by waterborne silt that had probably leaked into the cellar during its use life. I carefully troweled away the soil to expose the entire vessel, crushed into about 20 pieces. It appeared simply as if a saucer laying upside down on the floor had been stepped on by a boot.

After the fragments were cleaned in the lab, I began to reassemble the vessel. I was able to create two half-portions of the saucer, decorated in a common broad-brush, monochrome blue-floral pattern. But when I proceeded to glue those two final halves together, they would not fit. Upon closer inspection, the saucer was in fact two saucers. Both of which had originally been broken in almost exactly half, but both missing their respective halves (Fig. 2). They were nearly identical – probably from the same set – but upon closer examination the vessels were of slightly different diameter and the execution of the painted patterns was slightly different - reflecting the natural variation in hand-painted motifs. Somehow, two neatly broken halves from two saucers were laid together upside down on the floor of the cellar, as though they were a single vessel.

Fig. 2

Two pearlware saucer halves found on the floor of the Carmen Tavern Sangamo Town, Illinois

The circumstance behind such a find is utterly perplexing. And given the fact that the site was associated with Lincoln, the material from the cellar received a certain level of enhanced attention. Thus, we were on the lookout for a good story. This was the most interesting story in the cellar, and if one were to take the conspiratorial approach (as did some with the Umbrella Man), one might suggest that a mischievous and very foresighted individual was playing a trick on the archaeologists of the future. But of course, this interpretation is even less likely than a projectile fired from an open umbrella. Instead, the saucers will remain a conundrum – and a small one at that.

The dangers of a lurking “archaeological umbrella man” - an object that is not what it appears to be, or a kind of material non-sequitur unconnected to actions or events as it would seem – undermine our routine confidence in the fine-grained interpretation of seemingly well-understood objects. The whiskey flask found in the well of the Methodist preacher may not betray a drinking habit, the Staffordshire saucer may have nothing to do with the genteel practice of taking tea, and that iron kettle may have held something much less pleasant than Sunday supper. And as interesting as the appropriation of these mass-produced goods would be to the anthropologist, that renaming is generally invisible archaeologically. Thus, confidence in the basic building blocks of archaeological interpretation begins to fail if examined too closely.

Instead, the function-as-designed role of these objects generally holds up more reliably in aggregate – or, in a way, as seen from further away. Ten saucers from a frontier Illinois home suggest that someone is indeed using at least some (or most) of them for their intended purpose. Forty scroll flasks from 15 privy shafts in downtown St. Louis probably reflect neighborhood whiskey consumption – at least the subset that was discarded into the outhouse. So, sample size does two things – it allows for Occam’s Razor, and it may tend to “factor out” any lurking Umbrella Men such as the flask used for turpentine or kitchen kettle toilet.

The Fashioning of Artificial Umbrella Men

However, we can also introduce our own, artificial Umbrella Men as part of the study of these things. This can occur as a result of overzealous interpretation, often tied too closely to the meaning or significance of a single object, or from conclusions drawn from too-small samples. Essentially, overzealous logic can make for fictional outliers. Even more common is the problem of what my colleague Duane Esarey (pers. comm. 2011) described as “assumption stacking”:

This is something I mentally envision as the effects of gravity on stacked glacial cobbles. In archaeology, I've seen a tendency for a number of people to have pet theories which they envision that they have generated as proceeding from a single observation that leads to another and another, etc.

(Duane Esarey, personal communication 2011)

As an example, a study of an early twentieth-century domestic feature assemblage from Santa Monica, California, found remains of five porcelain or bisque doll heads, all of which were very fragmented (Wilkie 2000). Because no doll arms were present, and because the toys were more heavily fragmented than other fragile items in the sample, it was concluded that the dolls reflected not only the presence of children, but the action of intentional breakage. Further, it was argued that the intentional breakage was at the hand of a single child – a female. And further yet, it was surmised that the young girl may have intentionally broken five dolls as a statement to her parents. Synthesizing the archival record of the household, it was argued further that the birth of a second female in the household during the use-life of the feature was the cause of this breakage. That is, that a young girl objected to the presence of her new younger sister, and protested this change in her household position by breaking the dolls.

An impressive six assumptions were used – or stacked - to draw this intriguing conclusion. Firstly, was the character of breakage – that other fragile items should have been just as fragmentary as the doll heads if general random action had been applied evenly across all of the debris (which of course is generally not the case in most samples). Secondly, that the negative presence of doll arms contributed to the interpretation of the presence of doll heads. Thirdly, that the dolls were intentionally smashed, and at the hand of a child, who was female (no malicious boys around), and that the action was intended as a statement to others (as opposed to private action). Those assumptions led to the conclusion that a psychological trauma of a young girl in 1921 was visible in this archaeological assemblage.

While fascinating, the scenario also reminds one of the elaborate drawing of an umbrella-as-firearm schematic that was inspired by the presence of the man with the umbrella in Dallas. In this case, however, five broken doll heads were not an unusual presence (as was the man with the umbrella). Occam’s Razor would state that they logically reflect the presence of children on site, and their condition reflects the random actions played upon various primary and secondary discards over time. Thus, an Umbrella Man was created with stacked assumptions. The result was intriguing, but not supportable. Further, it suggested the promise of a level of interpretation that is really not available in the archaeological record. We probably just will not see such things, or more accurately, we will not recognize them or be able to support them if they actually happened.

Another instructive example demonstrates both the presence of an unusual circumstance, and also the precarious stacking of assumptions on the part of the researcher. This example comes from a 1980 study of an 1830s store ledger from the small town of Petersburg, Illinois (Kwedar et al. 1980). The study calculated the various types of goods sold at that store across seasons, months, and days of the week. Such information is often quite useful in better contextualizing the fragmentary remains of some of the durable goods found archaeologically.

In this case, in all of the pages of sales, only one sale made on a Sunday was recorded. Thus, it was logically assumed that the store was generally closed on Sundays. However, this unexpected outlier (like the man with the umbrella) then prompted a closer inspection of what was actually sold on that particular Sunday. The item in question was recorded as a “c.c. chamber.” The authors of the historical study, not familiar with certain commercial terms used by early nineteenth-century wholesalers and retailers, concluded that "c.c.” must have referred to the unit of measure (cubic centimeter) often used in modern medicine. Thus, it followed that the “chamber” involved must have been some sort of medical device or product. Further, the fact that it was sold on a Sunday suggested that this device represented the needs of a family emergency. That emergency prompted the storekeeper to open on a Sunday. The picture that emerged from the study, and from several stacked assumptions, was of a storekeeper graciously opening his store to a customer who was struggling with the illness of a family member or friend, and that this item could potentially attend to some sort of medical emergency.

In fact, “c.c.” was a common term in retail and wholesale sales for “cream-colored” ceramic. And “chamber” was the quite common abbreviation for a chamber pot, or the receptacle that served as an indoor toilet during the early nineteenth century. This presents a much different scenario, and presumably a much less dire circumstance. Occam’s Razor would tell us that, for some reason, the storekeeper was present on a Sunday, and a passerby took advantage of his presence to purchase a chamber pot. Sadly, for the literature, the reflection of a medical emergency on the Illinois frontier was in fact the reflection of the circumstantial purchase of a common household article on a day when a store was not normally open.

Degrees of Interpretive Distance: The Close View

To return to Updike's observations, its seems quite possible that his perspective on the Umbrella Man and the consequences for historical investigation is relevant to archaeology. That is, in looking too closely at our excavated data, we may find an unexpectedly disordered universe of interpretations and expectations. That the messy complications of idiosyncratic practice will frustrate the historian, and will also work against the construction of reliable patterns, unless sample sizes are large enough to obscure the outliers. And finally, that our own enthusiasm and natural attraction to the conundrum can potentially create its own umbrella men where there were none – such as the urgently needed cream-colored chamber pot in Petersburg.

We work in a discipline of assumptions, many of which become stacked. The “gravity” that threatens the stones of those stacks, in this case, is the unpredictability of human behavior when viewed in a micro scale. And yet it could be said that as archaeologists, on a site-by-site basis, we spend most of our time looking through figurative microscopes. We study the remnant and the fragment – the small percentage of a much larger whole that has decomposed or has been carried away from a site. Troweling the base of a pit cellar for a thumbnail-sized piece of a pearlware saucer is, at least initially, working at a micro scale. So perhaps we must remember to use caution, and not allow the proximity of procedure to influence the scope – or intimacy - of the information we seek.

There is an undeniable reality to the seemingly nonsensical factors behind any historical circumstance. Manufactured goods were regularized, but their use and disposal was not. And if one looks too close - when one attempts to maximize each detail from each artifact - probability, predictability, and indeed Occam’s Razor may fail. And unlike the Umbrella Man in Dallas, we will not be able to summon our subjects to testify as to the meaning of their actions and thus solve our small mysteries.

Of course, post-processualists some time ago encouraged the longer view toward pattern over the close focus on particular artifacts (e.g. Boivin 2008, Hodder 1982, Shanks and Tilley 1987). Associations, patterned relationships, and broader contexts not only offer new ways to consider meaning, but they also insulate our conclusions from stray Umbrella Men. In aggregate, probability will more safely rule the interpretations of past actions than will the single item. And we already know this, as any statistician would observe, reliability follows sample size. The gravity of reasonable definitions of ordinary things returns, but only if we pull back to a position that accommodates the occasional Umbrella Man, and yet still manages to speak of some sort of reasonable, probable truth.

When we concern ourselves strictly with history – the telling of local time – this path is straight enough. But when we claim to discern motive, choice, and practice (very personal phenomena even in the present) through the twice-fogged lens of time and decay, we enter a realm where the basic foundations of assumption could indeed be threatened by the inevitability of inexplicable umbrellas in every home and in every lifetime.

And this leads to what should be regarded, at least in historical archaeology, as a rather disturbing conclusion: that our most reliable truths will usually be terribly prosaic. The lot owned by the wagon maker produced more wagon parts (and here is what they look like). The wealthy merchant in the city owned more porcelains than the middling farmer out in the country (so archaeology can indeed detect a common truism about wealth).

So once again, what exactly are we looking for, or hoping to find? Can we somehow embrace the hidden, inexplicable interpretations behind any of our familiar objects? And how does that shape our confidence in the description of past lives? Can we reach beyond the probable circumstances that pass Occam’s Razor, to be inspired by rather than discouraged by the baffling disconnection between a black umbrella on a sunny day, situated at the very point of a completely singular historical event?

Degrees of Interpretive Distance: The Middle View

As Updike observed, that examination cannot be so close as to be distracted or misled by the Umbrella Man. However, our common position from the middle range may also produce its own form of distractions – in the form of patterns that lead nowhere. For the last 50 years in historical archaeology, the view from the aggregate middle has produced massive inventories of plates, cups, bottles, table knives, buttons, and smoking pipes that fail to provide much more than illustrations for a rich written record or support for prosaic truisms that we already hold (Mazrim 2019).

In Illinois, for example, an impressive body of archaeological data has been amassed. In 2006, the Illinois State Archaeological Survey initiated a web-based digital report archive consisting of approximately 24,000 files. The database includes survey and excavation reports from all 102 counties in Illinois. The archive includes various search functions that allow users to access reports by region, temporal affiliation, site name/number, or author.

In order to assess the nature of data contributions amassed in compliance-driven historical archaeological projects, the CRM Database was “mined” in 2016 (Mazrim 2019). All reports on historic sites that received Phase II or Phase III excavations were targeted. Some of these dated back to the 1950s, but comprehensive coverage begins during the 1980s. A total of 701 reports on sites or surveys that produced historic materials were downloaded from the CRM database.

From the total 701 reports examined in 2016, 331 did not actually contain excavation data, being survey reports, reports on pre-Columbian sites with stray historic materials, etc. From the remaining 370 reports (many of which contained the results of investigations on multiple sites), a total of 416 sites were recognized as historic occupations that received actual excavations.

After two months of reading so many reports, we found that we had learned more about the history of historical archaeology in the region then we had about the character of that archaeology. And even less about the character of the people who left those materials behind. We saw a discipline that attempted to evolve from methodologies used on pre-Columbian sites and that was suddenly faced with a complicated built environment and an enormous amount of debris from the age of industrialization, mass production, and mass marketing.

Most concerning was the fact that temporally discrete deposits were rarely recognized, much less separated, during analysis or presented separately in the reports. As a general practice, features or discrete artifact samples from specific periods were usually lumped together during the analysis and then reported as a single unit of data. From a database point of view, nearly half of the historic sites appear as a long-term aggregate of materials and subsurface disturbances, dating circa 1825–1900 and thus are of limited analytical value. These describe the nineteenth-century archaeological resources on most sites as a single temporal soup.

Perhaps less surprising in this compliance literature was a consistent lack of any form of research design, or at least, an attempt to standardize data sets to establish patterns within a site or to be compared to other sites. Before the mid-1990s, this was often due to an inability to properly quantify consumer goods. Fragment counts sufficed over vessel counts, thus creating not a picture of consumer behavior, but a picture of how badly damaged garbage became while laying on the ground surface. Only in a precious few instances were there attempts to establish patterning of any kind. Instead, these largely descriptive reports pictured the remains of a built environment as represented in the subsoil, and then dutifully counted fragments of pearlware, ironstone, bottle glass, and nails as if they were relative quantities of mineralogical specimens mined from the same source.

Looking back at the regional literature, and how the discipline evolved as it examined the materials from the middle-resolution of artifact counts and descriptions of brick-lined cellars or painted teacups, it becomes clear how certain research themes proposed early in the practice - such as settlement, subsistence, and architecture - have proven analytical dead ends. This was due primarily to inherent limitations in the archaeological record, and answered by the better-suited strengths of the written record. More recent themes, such as socioeconomic status or ethnicity, have also produced disappointing results (e.g. Mazrim 2002) - largely due to the inherent problems of inexpensive, mass-produced, and widely-distributed goods coupled again with the simple fact that many such questions can be explored much more eloquently from within the written record.

New themes, such as the visibility of consumer choice and consumerism, are very probably better tailored to the archaeological debris found on individual home sites dating to the last three-quarters of the nineteenth century. But it still remains to be seen now meaningful these insights will be. The frequent appearance of self-made Umbrella Men in the literature – such as the angry young girl smashing the heads of her dolls - suggests that if archaeologists are indeed recognizing patterns, they don’t make as compelling copy as do small narratives that may or may not actually be supportable from materials in the ground.

Degrees of Interpretive Distance: The Long View and the Noise

What about the view that might be found by stepping even further away from the bits and pieces? If we pull back for a longer view, with its attendant lower resolution, what do we see? Perhaps it has something to do with how the traces of practice found in archaeological remains may still reflect the accumulated noise of human experience at a particular place and time. Not the individual actions or choices, and perhaps not even reliable or meaningful pattern in many cases, but instead, a din of interconnected decisions and traditions that were really only articulated through inadvertent expression and random occurrences within an enormous material universe that was constructed by one purchase, and one choice, at a time. Over and over, across the county or down the city street. Could it be that what is contained in 50 years of gray literature and thousands of curation boxes is best described as a genuinely authentic noise? A record unlike that found in documents or photographs – less precise but somehow more authentic through its lack of self-consciousness? A record that can only be best appreciated at a precise distance that we have yet to ascertain?

I would suggest that this noise of experience will still have certain shapes and certain tones, perhaps based on temporal, geographical, and social circumstances (or perhaps based on something else) that can still be observed and described. This may be difficult to articulate through a series of Excel-driven pie charts based on typical classifications of mass-produced goods, such as those I currently include in each of my middle-range CRM technical reports. But I think an answer lies in Updike’s allusions to the laws of quantum physics and the nature of history. Specifically, concerning the distance from which we observe a particular subset of the universe (or time and place in history). That this distance dictates how we describe the order, characteristics, boundaries, of human experience as reflected by a fragmentary material world.

After 30 years of digging and reporting historic sites in the Midwest, but still without a convincing mechanism for the scientific support of this statement, I would observe that the site of a circa 1820 fur trader’s cabin in the French creole community of Prairie du Chien, Wisconsin, just feels different than the site of a hog farmer’s log cabin in rural central Illinois. And that a box of material from downtown St. Louis from the same era is immediately apparent as reflective of a much different human experience. There is an impression, composed of very real elements, which becomes more acute with experience. Sorensen (2015) has recently suggested the consideration of an archaeology of atmosphere when interpreting physical remains, but in that case, the approach is applied toward physical spaces as opposed to material assemblages.

In any case, once in the lab these impressions tend to break down, in part as one zooms in on specific items - many of which are often duplicated between various assemblages. If I look too close, I can also get lost in the specifics of the beaver trap from Wisconsin or the French cologne bottle from St. Louis. But human experience is hardly represented in a few items that tell time or speak of trade anyway. In aggregate, however, samples of debris from each of these worlds are, at least perceptually, different. “Perhaps, as with the elements of matter, investigation passes a threshold of common sense… where a rough kind of averaging out must substitute for the absolute truth” (Updike 1967).  The noise is just different.

The Attribute of Place

Its seems more than probable that one ingredient of this less quantitative impression is that of the sense of place – something relevant not only to the former occupants of a site but also to those who return there to dig. A sandy ridge overlooking the Upper Mississippi River is very different than the loess-covered expanses of former prairie on the flatlands of central Illinois, or the more complex built environments (both then and now) of downtown St. Louis. Again, that sense of place as used here is still an impressionistic one, and not necessarily akin to landscape-based perspectives found in emerging studies of the “archaeology of place” (e.g., Bowser 2004). Whitridge (2004) has observed that the former need not solely define sense of place in archaeology, and that place can be composed of “a nexus of imaginary significations” that when considered, open the door to “hybrid past realities” (Whitridge 2004: 214).

In 1825, a painted Staffordshire saucer such as those I found at Sangamo Town had a certain meaning to the residents of that village, based not only on who they were or what they needed a saucer for, but also their understanding of their position in the world that produced such things. Their role in the world they imagined beyond the horizon was mediated by how they regarded objects that crossed those boundaries. The meaning of a certain object was (and still is) different in different places. Nothing was fixed (e.g., Bronner 1985; Johnson 1996; Miller 1987), and it was conditional to a variety of circumstances - including time and place.

Take the case of Madeline Island, Wisconsin. The island is located 2 mi (3.2 km) from the northern shore of Wisconsin in Lake Superior. It is and has always been a sparsely occupied place – due in large part to the harsh winters and an environment that hinders opportunity. During the eighteenth and early nineteenth centuries, those who lived on the island - Ojibwe and their recent French-Canadian partners – engaged in fishing and the fur trade to make a living. Then, as now, the community was a remote one.

On the south end of the island was the compound of the fur trader Michel Cadotte – a Canadian Metis (Ojibwe) trader who established a remote post in the far western Great Lakes (Ross 1960). The character of the place was well summarized by a visiting American in 1826:

On nearing the shore, (which was grateful to my feelings beyond the power of language to express; for it looked green, and had the evidences of civilized life …

The first question I asked on landing, was to know of Mr. Cadotte, who has lived here twenty-five years, if he had any milk, and was rejoiced to get the answer “Oui Monsieur.” I never enjoyed this article before. It tasted like nectar - and I thought I should never get enough of it. His houses were thrown open for us, and all he had was put freely at our disposal…This is the only spot that has brought gladness to my heart, the associations of home and of civilized society, during a voyage of four hundred miles - since we left the Sault. It looks like a fairy scene, and everything about it is enchantment. Yet the houses are of logs; but are lathed and plastered.

Thomas McKinney 1959 [1826]

Excavated at the site of a compound (and displayed in a local museum) are a range of items associated with the fur trade and early nineteenth-century domestic activities – including fragments of several blue-painted saucers similar to those found in the tavern at Sangamo Town. They were made in the hundreds of thousands during the 1820s, and I encounter them everywhere. In this case, these manufactured consumer goods followed a rather different trajectory than those I am used to seeing on Midwestern farmsteads or in the small towns and cities near the central Mississippi River Valley. In the latter, such saucers were common, everyday commodities – parts of generic “crates for the country trade” (e.g., Ewins 1998; Mazrim and Walthall 2002) sent out to retailers across the new settlements. They were ubiquitous, easily had, and probably only noticeable when they were absent from a typical farm-family household.

When I unearth such saucers in central Illinois, I know full well how they got there - as did their consumers. Carted out from busy urban entrepôts such as St. Louis to hundreds of retail country stores. They were the subject of exchanges made by people of the same general perspectives and histories, focused on the comfortable fashion of then-modern, middling goods.

The journey that the saucers made to wind up buried in the sand on Madeline island was quite different. I am not quite sure where Cadotte got his saucers, and I doubt he saw them the same way as did the tavern keeper in Sangamo Town, 600 mi (965.6 km) to the south. The trade would have come from the Northeast, and the saucers were probably packed into a canoe from Montreal. And that canoe would not have been laden with the same kinds of goods that filled the flat boats and steamboats plying the central Mississippi Valley. Instead, most of what Cadotte unpacked on the beach consisted of goods for the fur trade. And as a merchant, Cadotte was conscious not so much of European fashion as he was the tastes of old Native cultures and new Metis traditions in the north woods.

The blue-floral imagery painted across the surface of those sauces actually reflected the recent transition in Staffordshire factories away from ancient Chinese imagery and toward broader designs more reminiscent of more recent Dutch traditions. However, the merchants and consumers of those saucers across North America generally knew nothing of these things. The imagery was popular and ubiquitous during the 1820s. However, those motifs very probably meant something different in the tavern at Sangamo Town than they did at Cadotte's post on Lake Superior. In fact, the painted patterns on early nineteenth-century pearlware just so happens to be aesthetically complimentary to the traditional beaded designs that decorated the clothing and accessories of the Ojibwe and Metis residents of Madeline Island. It may be for this reason that one such saucer (Fig. 3) was found in an early nineteenth-century grave on the south end of the island, resting upside down in a brass kettle filled with wild rice. It had not been used for tea, it was instead used as an appropriately decorated bowl for rice by someone who probably spoke only Ojibwe and French. And then it became part of the provisions of the dead.

Fig. 3

Painted pearlware saucer from the Cadotte Site on Madeline Island, Wisconsin

Much of this was apparent to me when I first saw the specimens unearthed 200 years later on the shore of Lake Superior. The saucers from Sangamo Town and the saucers from the trading post were of the same design but, because of where they were used and buried, they had developed very different meanings. And on the island, they had been mixed with other objects of similar mismatched trajectories, creating a general aura of “displacement” on that wintry beach that cannot help but shape how the object is regarded, both then and now. Perceptually, the saucers from both places have acquired what in the wine trade is known as terroir, or the manner in which the climate and soils of a particular region affect grapes in such a way as to impart distinctive flavors to the wine.

The teawares from Madeline Island were also not unlike the man with the black umbrella standing out in the sun in Dallas – out of place and not what they appeared to be. But unlike the Umbrella Man, their meaning is decipherable - if one knows something of the history of the place, and if one steps back.

The Umbrella Man is a powerful example of the fundamentally nonlinear character of human experience. The circumstance of 1963 demonstrates, in a particularly acute manner, that there is no reliable way to place so many human gestures, choices, or experiences into categories or patterns that are based on linear, analytical deduction. The circumstances as they appeared on the surface on that sunny day in Dallas were completely misleading and utterly undefinable to the observer. The Umbrella Man, I believe, is in fact a representation of a greater truism - that linear methods cannot be successfully applied to nonlinear phenomena. In mathematical terms, applying a linear trend line to a complex, non-linear (and sometimes unreliable) data set will produce a deceptively simplistic pattern. The “standard deviation” in the dataset that is human action will ultimately be too great to support a reliable “R-squared value” (Shapiro and Wilk 1965).

In 1996, Charles Orser used the term “multiscalar” to describe an archaeological approach that considers analysis at differing scales, not unlike the “degrees of interpretive distance” discussed above. He too recognized the inevitable problems of contradictory evidence from the “entangled past” that are illuminated by scalar approaches (Orser 1996: 186-187). However, Orser suggests that there is indeed some possible level of analytical “transcendence” achieved through the step-by-step application of a multiscalar analysis. In other words, instead of abandoning the lower or middle-resolution perspectives described above, it may simply be necessary to pass through each of these stages to ultimately best describe the “netlike connections” (Orser 1996: 187) of human experience as reflected by archaeological debris. An appropriate analogy, then, to the “degrees of interpretive distance” or the “multiscalar approach” would be the three-lens compound microscope. And no single lens will be sufficient to describe the subject.

The practice of history represents our natural impulse to define linear patterns from our daily nonlinear realities. The Umbrella Man is a wonderfully neat and clean example of the messiness of human action and historical interpretation. However, as individual human experience itself is not linear, linear historical perspectives will probably ultimately fail to define and characterize that experience at all but the softest (or most distant) of resolutions. And of course, that experience is not defined by the superficial meanings of mass-produced goods. Those meanings are unintentionally altered by temporal, geographical contexts and by the individual agencies of appropriation and renaming and are very difficult to quantity or discern from an archaeological distance. Either the Umbrella Man phenomena can be read as disheartening, or it can be regarded as an inspiration. Perhaps it can motivate us to find another position (or multiple positions) from which to observe the fragmentary past, and to develop non-linear methodologies that are somehow more analogous to human experience. Adjusting our lenses, our expectations, and our descriptive tools, to the noise.

Meanwhile, the two halved saucers from Sangamo Town remind me that while ceramics are largely non-perishable – the circumstances behind their use and deposit into the earth are far from it. The saucers were designed for tea, were probably used for some sort of beverage service (or perhaps one of them as a grease lamp), and take their place comfortably in a large assemblage of commonly occurring items found on early nineteenth-century sites. However, the circumstances that led to the two equal halves laying on the floor next to each other and in isolation, however perplexing, cannot be determined. And further, even if those circumstances were to be revealed, they would probably offer little of significance to the understanding of daily life on the American frontier. Instead, the two saucer halves probably reflect the not-very-interesting mysteries of debris deposition. However, when pulling back a bit, they also wink at more interesting questions concerning the impetus behind, and the cultural uses of, the variable-focus instrument that is the archaeological endeavor.