Working knowledge

The time has now come to pick up the track laid down earlier, about the profound and numerous individual differences that divide us. A world of people with different personalities, different aesthetics, different interests, different levels of drive and motivation, different life experiences, and different problems is not going to be one of perfect cultural sharing. This explains the diversity we have noted above.

So a society must be built from highly disparate individuals, and a culture formed from highly disparate knowledge and interest pools. Highly disparate emotions motivate and drive the process.

Only mirror cells and the desperate need to conform hold us together. This leads, in societies like modern America, to the “lowest common denominator” phenomenon. People cling desperately to the most idiotic manifestations of culture—sports, “celebs,” pop music, TV series—because only these can be learned easily enough to become common ground. The bigger and more diverse the culture, other things being equal, the more idiotic the popular culture. This is not always true, but it has enough rationale to make it a general law. Exceptions are marked cases and must be explained.

Cultural Mixing is the Norm, Not a Marked Case

This sort of thing has recently produced the truly ridiculous concept of “cultural hybridity.” This idiotic bit of racism implies that cultures (and presumably their bearers too) are separate species, that can “hybridize” like horses and donkeys. In fact, cultures, worldwide, are all part of one great interlocked network. In a very real and literal sense, there is only one culture in the world: human culture. It has been one ever since humans emerged hundreds of thousands of years ago, and certainly since modern humans radiated from northeast Africa—an area small enough to be a single cultural realm—a mere 100,000 years ago (give or take a bit). Hence the interlocking, overlapping, and cross-cutting systems of thought that in fact characterize culture worldwide.

Even a blend of really disparate cultures, like the syncretism of pre-Columbian and Spanish Catholic religions so common in Mexico, is no hybrid. It is a re-mixing of the waters in a vast braided river. The river is the same, the water is the same, however much the channels may diverge and reconnect. Or, to use the image A. L. Kroeber used on the endpapers of his famous textbook of anthropology (1948), the tree of culture differs from the family trees of biology in that the branches always keep regrafting back onto each other.

Key inventions, like fire, pottery, the bow and arrow, agriculture, and more recently automobiles and computers, spread all over the world. They are cultural, but now they are pan-human. Even the most isolated hunter-gatherer bands are now familiar with trucks and are apt to be online with other hunters half the world away. The idea of monolithic, isolated “cultures” thus has nothing to do with reality.

Ralph Linton’s famous and oft-reprinted essay “The One Hundred Per Cent American” (Linton 1937) is easily found now on the Internet (often mis-cited!), and needs repeated attention. Linton humorously pointed out how the militantly patriotic American draws on the entire world for articles of daily living—almost nothing actually stemming from America itself.

Think of what an anthropologist could confidently predict about a new society found on a newly discovered island.

We would know the people had an elaborate, complex language, with many sounds, with nouns and verbs, and indeed with enough humanwide traits to fill a dense book, Universals of Language edited by Joseph Greenberg (1978). No animal has anything remotely close. We would know the culture had dozens of humanity-wide culture traits as well, and here the book is Human Universals by Donald Brown (1991).

We would know there were rich resources of language, art, music, oral literature, dance, ritual, and food preparation. We would know these were often put in the service of a complex and abstract religion characterized by belief in supramundane (“supernatural”) beings, in human ability to get power from the supramundane realm, and in human ability to use that power to heal illness and bring better luck in weather, fertility, or similar matters.

We would know the group had a complex, formal family and kinship system, involving rules of descent, inheritance, and marriage. It would have a political system of some kind with at least some hierarchy (if only a tendency to defer to elders), a system of gender roles, and a system of unwritten laws and rules with some enforcement mechanism (Abrutyn 2009). We would know the group had some sort of economic system, if only one governing exchanges; trade would take place in a highly structured and culturally embedded way, not as a casual procedure.

We would know the group had a moral system. It would idealize generosity, self-sacrifice, solidarity with the group, and following the group’s rules. It would usually praise even-tempered, accommodating behavior, but would have an escape hatch allowing its members to stick up for their rights even if that meant some fiery dialogues. (It would not, however, have a predictable list of what those rights were.) It would negatively judge disruptiveness, random violence, robbery (but not always theft), stinginess, treachery, and other social disruptions.

If the society were large (say 100,000 people in a polity) and complex, we would strongly suspect it had a written language and had written down its major laws and economic transactions. Such a large society would certainly have monumental architecture, a ruling class, a priesthood of some kind, and some sort of redistribution mechanism such as taxes or tribute. It would have to have storage facilities and record-keeping to manage that. The amazing parallels between all early civilizations, including the Native New World ones that were completely isolated from Old World influence, shows that these are built-in and necessary features of any large cultural group.
Culture as Distributed Cognition

Cultural knowledge is most sharply and clearly defined at two ends: what all members of a culture learn, and what only experts know. Some anthropologists (especially of an earlier generation) regard culture as the common knowledge of a group: the knowledge that every normal adult shares. Others (more typically of my generation and younger) take the sum of the knowledge held within a group, including even the esoteric material known only to a few.

Most cultural knowledge actually occupies the vast space between these extremes. It is the knowledge that some or most people have, but that is ambiguous, loosely structured, loosely enforced. All children know basic English grammar: “s” for plurals, “-ed” for past tense…. Professional writers know publishers’ stylesheets by heart. In between is a vast realm of loosely consistent rules that most people “sort of know.” Is the subjunctive dead? When do you use a semicolon? Similarly, most American males of my generation know something about fixing cars, but many don’t, and everyone who does know seems to have a different knowledge pool; some know Toyotas, some Fords; some do brakes, some are best with radiators. Clearly, this is not the precise, systematized knowledge that a professional mechanic has. The idea of culture as 100% shared knowledge cannot stand up.

In short, culture is distributed knowledge (Kronenfeld 2008b). Even the simplest society has its experts on hunting, on medicine, on plant foods, on religious lore, and so on. A modern society has millions of members, among whom every adult has a unique knowledge pool. Measures of cultural consensus (Romney et al. 1986) are valuable precisely because consensus is unusual. One expects variation.

David Kronenfeld (2008b) has even made this the basis of a theory of culture. Humans being each different, and knowing different things, they require some basic sharing. Distributed knowledge requires coordination. Different members of society have to know who is doing what, when, how. “[C]ulture provides us with the information that we need to be able to play our role in the larger social enterprise…. Differing levels of shared knowledge enable differing levels of subgoals and subtasks.... [This] allows us to communicate and coordinate with others….” (Kronenfeld 2008:94).

Even in a single tightly-knit linguistic community—say, a small, traditional English village—there are enormous differences in knowledge of vocabulary, talkativeness, and usage. The blacksmith, the sheep farmer and the vegetable farmer each has a specialized technical vocabulary. Language is what is culturally known; speech is individually produced. Culture can tell us to say certain things ("Hello, how are you?") but can never determine the exact content of a long dialogue. Every conversation is shaped by personality, family role, life experience, immediate concerns, and immediate events, as well as by cultural expectations.

People try hard to conform to cultural rules, but no two people interpret the rules in exactly the same way, and no two people face the same situations. Everybody has, in his or her head, a different (and constantly changing) version of the "common" culture. Brilliant approximators that we are, we can always find common ground, and describe our culture in generally accurate ways. This is one of the many reasons why “memes” do not and cannot exist. Uniformity of cultural units is limited to visible, tangible things like arrow design, in which people can copy each other faithfully.

People learn categories, shortcuts, and other ways to simplify the world. We could not possibly manage the full complexity of the world without those. But, without considerable care, categories become barriers to thought. This can lead to serious barriers: social conformity for its own sake, and defensive rigidity. Humans need to find ways to de-barrier their thinking, especially as they get older. Mystical meditation helps. It enables one to leave the cultural categories, but not into a world of chaotic, unfiltered stimuli. Cultural categories disappear. Chinese and Tibetan meditative training teaches this rigorously.

Overlearning cultural “truths” is a constant danger, and any thinker needs to worry about it often. I find valuable, even necessary, an instruction ascribed (by folklore) to George Bernard Shaw: Every morning, for mental exercise, he tried to argue himself out of one deeply cherished belief. I have no idea if he really did, but if he didn’t, he should have. So should we all.

The fact that such discipline is necessary proves how difficult it is to break out of culturally learned ways, even when these are totally inappropriate for the situation. The long and unedifying history of explorers who refused to learn from local people, and thus died from eating poisonous foods or wearing hopelessly inadequate clothing, is most revealing in this regard. It was traps like this that led early anthropologists to see culture as a straitjacket. The key point, however, is that people do not need to trap themselves so thoroughly. Many explorers did listen and learn. The non-learners were more animated by prejudice than trapped by culture.

Often, though, it does not matter a great deal which “explanatory model” (Kleinman 1980) the healer holds. People around the world have learned that willow tea dramatically reduces fever and inflammation. It makes no consequential difference in the treatment whether this action is explained by God’s blessing to suffering humanity, by the action of the Spirit of the Willow Tree, by a Galenic cooling principle in the tea, or by the chemical action of methyl salicylate on prostaglandin metabolism. Straightforward reality is quite enough. Everybody knows that willow tea is effective. (We now use its Bayer-created derivative, aspirin. Aspirin was actually isolated not from willow but from another salicylate-rich plant, formerly called Spiraea, of which “aspirin” is an anagram.)

Chunhuhub has its expert beekeeper, its expert hunters, its all-round expert folk biologists (like Don Felix), and its expert farmers. The leading expert healer is Don José Cauich Canul (see Anderson 2003) in the nearby town of Polyuc, but many in Chunhuhub know hundreds of medicinal herbs. Conversely, there are many people who are quite ignorant of local plants and animals, knowing only the domesticated ones and a few healing herbs and major trees. (Even so, these make up a list of easily 100 species in even a nonspecialist town-dweller’s repertoire.)

In Hong Kong I found similar contrasts. Some of the fishermen I knew could identify hundreds of marine life forms. These individuals were well known and highly regarded. Others knew only a few dozen. Nearby farmers knew even fewer, though even they could reel off three or four dozen fish names. The farmers knew hundreds of agricultural terms, mostly unknown to the fishermen. Interest, native ability, and family and peer reinforcement explain such differences. Some people who were not especially knowledgeable about fish were experts on kinship, or religion, or healing.

To be learned, knowledge must also be available. Esoteric knowledge is often unavailable to any but the initiated. A huge percentage of important ecological knowledge in aboriginal Australia falls into this category; it is sacred, and taught only to initiated adults. Some is strictly men’s knowledge, some strictly women’s. Healing knowledge among the Yucatec Maya includes many “secrets,” passed down only to those accepted as students by the hmeen (healers and ceremonialists); most of these are the hmeen’s children or other close younger relatives. Almost all societies except the very simplest have institutionalized esotericism of this kind. Of course, such knowledge is generally prestigious. Outsiders are motivated to learn it simply because it is secret.

Also, many societies have equivalents to our copyright. Northwest Coast songs, stories, art motifs, and so forth are owned by individuals or (more often) families, just as ours are in the modern international media. Classical China did not recognize copyright per se, but did condemn plagiarism. The Maya have no such rules, but some knowledge is known to be basically family-held, and this is respected.

“Cultural models” have been variously defined (e.g. d’Andrade 1995; Kronenfeld 1996, 2008a, 2008b; Holland and Quinn 1987; Shore 1996; Strauss and Quinn 1997; Strauss 2006 has also contrasted the cultural model with the trendy idea of “the imaginary,” to the disadvantage of the latter). Psychologists, similarly, speak of cultural or sociocultural models (Markus and Hamedani 2007; see esp. their very comprehensive list of references, p. 16). Depending on the writer, cultural models can include such cultural constructions of knowledge as belief systems, institutions, institutional culture, priority lists, guidelines, rules (usually rules of thumb), canonical action plans, static plans (like standard floor plans for houses), and goal sets. They are the cultural equivalent of the “schemas” of psychology, a term also defined variously by various scholars. Depending on who is using the term, a cultural model is a set of beliefs that sort together and are clearly related in people’s minds, or, more specifically, strongly structured systems of knowledge that become traditional and canonical in a culture.

Specific concepts for objects and acts: “dog,” “walk”

Broad general concepts and concept-classes: ranches, airplanes, stereotypes of ethnic groups, concept of potato salad, ideas of a stylish coat vs. a hopelessly out-of-fashion one

Landscapes, ecologies, scenes, frames, other broad visuospatial representations: Yosemite National Park, a typical farm, an ideal country house, a slum

Classification systems: kinship, vegetables, kitchenware, cars

Texts, poems, songs, tunes, other linear communication forms

Stories, story lines, canonical plots (see Vladimir Propp’s classic study of cultural rules for creating folktales, 1968)

General ideas, vague ideas, overall views, other high-level and unspecific representations: human nature, animals, the real world, religion, Christmas

Associations and networks: my connections with my friends and enemies, the relations between the nations of Europe, the tension between sects of Christianity, the relationship of stars and energy flows in the galaxy

Canonical plans and action sequences: Schank and Abelson’s famous restaurant script (Schank 1977), directions for assembling furniture

General rules and rules of thumb: rules for driving cars, training horses, brushing one’s teeth, fixing up the house

Recipes. Cooking recipes are an excellent model for cultural knowledge transmission. Michael O’Brien points out that recipes tell “how, when, where, and why to produce something…[and] contain multiple parts of two general kinds—ingredients and rules—that can be reconfigured to form a different recipe” (O’Brien 2008:43).

Broad classes of behaviors with increasingly vague or abstract canonical rules as one goes up the recursion ladder: dating, marriage, and romance (Strauss and Quinn 1997; Victor de Munck, pers. comm.), religious rituals (Shore 1996), games (Wittgenstein)

Meta-representations: our shared ideas about others’ ideas and how others think, our ideas about thought and mind in general, our cultural concepts of culture. Every culture has its “theory of mind.”
This is impossibly broad, and strips the term “cultural model” of any real meaning. A model should be a simpler representation of a big complex thing, as in model planes, which can vary from a boy’s little lath toy to an aeronautic engineer’s complex mechanism, or even a computerized mathematical representation of that (a model of a model). The common thread is that a model is self-contained—you can set a clear boundary around it—and it is a simpler representation of something. By definition, all mental representations are simpler representations of things in the real world. (But poems, songs, stories, and the like are things in themselves—not just representations—and thus are not cultural models. A set of rules for writing a folktale, or a proper Arabic or Kaluli song, is a cultural model.)

The real problem comes in the “bounded and self-contained” part of the definition. Unbounded/open-ended systems and networks should not qualify. Yet, in the real world, no category is really bounded. Rules for driving grade into rules for bicycling, songs grade into chants and then into prose, and houses grade into stores. Thus, a model, cultural or otherwise, must be more bounded than the real world; setting hard boundaries on a vaguely-bounded conceptual space is the most important simplification that modeling performs.

One noteworthy way of tightly defining “cultural model” has been advanced by David Kronenfeld. He sees cultural models as “scenarios or action plans for how to behave in some given situation or how to interpret the behavior of others in one or another situation” (Kronenfeld 2008a:69). They contrast with organized knowledge such as classification systems, “’cultural modes of thought’ which provide metaplans for how to break down and organize an unfamiliar problem” (ibid.), and other knowledge systems. Cultural models would then seem to be a lot like grammar rules, and indeed linguistic research on grammar has had its influence here. Grammar rules tell you how to talk—how to construct a sentence that others will be able to decode by using the same rules. Cultural models would be largely procedural knowledge, drawing on discursive knowledge only as the action plan required. Structured systems of discursive knowledge, like folk taxonomies and lists of gods, do not qualify.

Such cultural models can often be broken down into flowcharts and decision trees. These are psychologically compelling. An adequate decision tree of a cultural process (such as planting a garden) specifies the normal steps taken, in clear and specified order. It lists the choices possible at each step, and the effects of alternative choices—branches in the decision tree. It directs the user to call up particular bodies of discursive knowledge at the right points. Such charts make excellent anthropologists’ models of the people’s own cultural models (Gladwin 1989; Hedrick 2007; Randall 1977; Young and Garro 1994).

The best models-of-models might combine the features of both recipes and decision trees. They would be set up like a recipe, but would treat the rules as progressive decision nodes rather than progressive cut-and-dried steps. Experienced cooks rarely make a recipe exactly the way the book tells them. They improvise, treating the steps as suggestions or decision nodes rather than as absolute rules. The recipe is the cultural model; each actual cooked version of it is a real-world instantiation that goes beyond the model and adds individual “spice.” The same is true of decisions in fishing (Randall 1977), ranching (Hedrick 2007), and other areas; people bring enormous amounts of knowledge to bear on particular decisions, and have a fine-grained spread of decisions, each one involving options. Language is the same: we have grammatical rules and canonical sentence forms in our heads, but we do not always follow them, and we construct new sentences all the time.

True models overdefine, oversharpen, make conscious much that is really preattentive, and in general make a clean, sharable process out of a messy, personal one (Young and Garro 1994). We are describing for analysis, not producing Jose Luis Borges’ worthless l:l map of the world (Latour 2005). A recipe for bread is a cultural model; it tells you what you need to know. A recipe that described every single baker’s personal touch would be millions of pages long. Scottish folksongs are composed according to a few simple rules—a cultural model—but listing all the songs would be impossible. New ones are being written faster than one could list. What we want of a cultural model is what its users want: something they can learn that will then let them personally improvise.

Since we have raised bread (the pun is inescapable), let us consider it. I learned to make bread partly from recipes and partly from people around me. I can easily reduce my bread recipes to simple half-page instructions. Naturally, I do not put everything I know into such instructions. I leave out the variations I often make. More importantly, I can’t put in the feel: the fingertip knowledge of how wet the flour is, how long to knead the dough, and so forth. It is too difficult to verbalize, and is largely an unconscious process anyway. One must first learn simple straightforward rules, then do much actual practice to get the feel, and then go on to learn more and more cultural and personal tricks. Breadmaking involves a cultural model which is then progressively personalized.

Kimberly Hedrick (pers. comm. email of Nov. 28, 2008; see Hedrick 2007) says of cultural models: “I see them as plans for action and for interpreting/explaining others' actions, and for predicting others' actions. In fact, I mostly think they're around to make us feel like our lives are more predictable than they actually are. It's stressful to actually acknowledge the divesity of what could be—both our own possibilities and then all those of others, plus how the world might actually work. It feels more secure (and streamlines our thought processes) to have mental outlines of what we expect from others, from ourselves, from the world. Of course, all of the above do stuff that are not aligned with the model, but then we can always extend the models in interesting ways so that we still feel secure. I think models exist as a way to integrate us with the complexity of social life (and life in the wilds of nature) in a way that maximizes cognitive efficiency (though naturally limiting cognitive accuracy) and lowers our stress about the unknown.”