Cultural transmission theory..also for research about web artifacts?September 27, 2008
Recently i found one paper elaborating Cultural transmission theory (CT) research in the anthropological sciences and outlining the beneﬁts and drawbacks of this theoretical framework for the study of material culture. This framework seems quite interesting for explaining the phenomena in artifact culture of the web communities as well.
Most interesting aspects for me are related with my ideas of ‘learning’ in niches:
- culture constitutes a second (in addition to genes) mechanism by which inheritance occurs
This claim is well related to the niche construction ideas from other studies and the evolutionary effect of niche construction.
- our analyses should include more than one measure of relatedness of cultural patterns, each valid but representing different information pathways.
This assumption makes me think of the various gradients that define niches, and how these gradients influence each other.
- humans link the difﬁcult information to things they can easily transmit as part of packages, and certain kinds of information will be bundled during transmission.
Is it the ‘view’ of affordances at some cases, they are somewhat externalized into cultural artifacts meaningful for the other actors coming mainly from this context?
Cultural Transmission Theory and the Archaeological Record: Providing Context to Understanding Variation and Temporal Changes in Material Culture
Jelmer W. Eerkens and Carl P. Lipo
J Archaeol Res (2007) 15:239–274
Explaining similarity and change in artifacts over time has been a long-standing goal of archaeologists.
Culture historians were interested in how these sequences varied from place to place and over time.
New focus now includes explaining other aspects about artifact variability beyond typology, including explaining why artifacts change the way they do.
Culture transmission (hereafter, CT) is simply the idea that similarity in behavior and artifacts may be caused by the exchange of information using a nongenetic mechanism.
Common descent in CT merely refers to the notion that information about material culture is passed between individuals and that similarity in artifact form may be a product of information ultimately coming from the same source.
Similarity in artifact forms over time and space was commonly explained by reference to the diffusion of ideas and information.
Diffusion was considered a general mechanism by which information was passed (or inherited) across and between populations.
Early and mid-20th century diffusion models were focused on the ‘‘culture’’ as a unit of study, and ideas were perceived as being diffused in and out of groups of people who comprise sets of bounded entities.
CT today derives from Darwinian models of evolution.
CT consists of the recognition that culture constitutes a second (in addition to genes) mechanism by which inheritance occurs.
CT theory provides a powerful means for linking measures of behavioral similarity and claims about historical relatedness.
Darwinian theory, of which modern CT is a part, is based more on the actions and decisions of individuals.
CT acts to decouple information transfer from biological reproduction and allows information to be continually passed from one organism to another through social learning.
The results of individual learning (i.e., behavior modiﬁcation) can be transmitted, in the modiﬁed state, to other individuals. Through individual learning and CT, organisms can continually acquire, modify, and pass on modiﬁed information. Thus, the process of CT is fundamentally based on the interaction of both individual experimentation (i.e., innovation) and social learning (i.e., copying).
CT can create patterns in behavioral traits that are distinct from behaviors controlled and transmitted genetically.
Cultural information may consist of a single trait from a single individual, the average of a trait in a group of individuals, the modal trait in a population, or any other combination from any set of models.
We also can acquire information as traits, sets of traits, or simply as rules on how to acquire additional traits or rules.
Behaviors transmitted culturally have the potential to evolve (i.e., change) quicker than those passed on genetically because there are few limits to the structure of information inheritance in CT.
Multiple patterns will generally characterize CT.
When we study cultural variability, our analyses should include more than one measure of relatedness, each valid but representing different information pathways.
As a result, in some cases cultural variability may lack distinct groups with clear boundaries and cohesive internal information, though such groups can clearly form.
Not all similarity in cultural behavior necessarily indicates historical relatedness.
Distinguishing between instances of historical relatedness and convergence will form an important realm in future evolutionary studies.
Where and how replication takes place in CT?
The lack of attention in this area leads to fairly simplistic notions of traits moving from individual to individual with frequencies that are driven only by their prevalence in the population.
Gabora (2004), notes that the locus of cultural replication is in the minds of individuals. Minds are more than simple ‘‘bags’’ that hold traits but complex webs of algorithms and rules for acquiring and especially sorting information. Gabora calls these algorithms a worldview.
As a person receives cultural information, it is ﬁltered through their worldview, where it is assimilated and related to all existing information before being stored and later recalled. Although strongly shaping the kinds and rates of information acquisition, worldviews are not static entities but constantly change.
The worldview not only transforms incoming information but is transformed itself to accommodate new information and is itself transmitted culturally.
First, we expect that individuals living in similar cultural, social, and physical environments will
tend to acquire similar worldviews. As a result they also may acquire similar kinds of behavioral traits, including material culture.
CT may include not only information about traits but also rules about how/when to acquire traits as well as rules about how/when to acquire new rules. This means that information may change (i.e., evolve) within a population at dramatically different rates.
Third, we expect that the set of rules that compose the ‘‘worldview’’ will be cumulative because they build on one another.
CT is a separate inheritance system that governs behavior, CT can easily account for seemingly ‘‘maladaptive’’ traits that spread or even come to dominate within the range of things people do.
For example, despite fairly strong genetically controlled instincts to eat, CT can explain why a behavior like anorexia may spread within a population (e.g., worldviews pertaining to a certain body image).
A signiﬁcant amount of information is preserved in some way and is coherently passed from individual to individual through populations over relatively long periods of time. How well the historical signal of cultural transmission is preserved, particularly over long periods of time, is unknown and surely varies from context to context.
We assume that at least some information is transmitted between individuals and that this information is subject to modiﬁcation before being retransmitted to others either through copying error, ﬁltration through worldviews, or purposeful ‘‘innovation.’’ As a result, the information that is transmitted is subject to evolutionary forces.
We suggest further that at least some of this information stays relatively intact over archaeologically relevant periods of time.
‘‘Information’’ that is argued to be transmitted between individuals: Unlike DNA, which is physically passed from person to person in Genetic Transmission (GT), no such empirical entity is known for CT.
We have no direct way of ‘‘seeing’’ transmission.
There is no physical ‘‘chunk’’ of material that is passed from individual to individual.
Fundamentally, there is no distinction between GT and CT; each system simply passes information using different ways of coding.
This information-centered view has several consequences.
1. There are no boundaries on the types of physical entity that can carry information. This is true for
cultural and genetic forms of transmission.
The lack of a single empirical entity means that we have to deﬁne one.
Although there are no agreed-upon empirical units of CT, we can build a unit for measuring CT. For example, Pocklington and Best (1997, p. 81) deﬁne CT units as ‘‘the largest units of socially transmitted information that reliably and repeatedly withstand transmission.’’ This deﬁnition makes it clear that CT units are measurements of the effect of transmission on variability, not a physical package of something.
2. We must keep the physical package separate from the information being transmitted. We are not interested in the physical package or set of physical packages of cultural information but rather the structure, content, and ultimate effect on observable phenomena like material culture.
Because variation is the raw material upon which evolution operates and CT processes directly affect variation, it stands to reason that CT strongly affects the course of evolution.
Transmission works on both the originating (i.e., source of information) as well as the recipient
side (i.e., the destination of information) as individuals acquire, store, recall, replicate, and materialize this information.
Different combinations of the content, context, the number of people involved, the direction of transmission, biases, and information packaging present a bewildering array of possibilities for CT.
Content refers to the actual information that is being transmitted between individuals.
The more complex information is, the longer it takes to describe its properties whether done
mathematically, pictorially, or verbally. For evolutionary modeling, the complexity of information is important because complex information is subject to greater copying error.
The various human sensory systems are different in their accuracy, hence the propensity to produce error during replication of cultural information (see Eerkens 2000; Eerkens and Bettinger 2001; Eerkens and Lipo 2005).
The repetitiveness of the information being transferred also affects error rates during replication. Information that is highly repetitive is more likely to be materialized with less error than information that is singular.
The structure of information affects how it is transmitted. Mesoudi and Whiten (2004) showed that social information loses detail (‘‘low-level information’’) but may gain high-level structure as it is transmitted between people verbally. Mesoudi (Mesoudi et al. 2006) suggests that social gossip is transmitted with greater accuracy than similarly structured but nonsocial information.
Washburn (2001) found that the overall structure of the images was more accurately reproduced than elements about detail. Furthermore, cultural background played an important role in the accuracy of reproduction; the greater the familiarity of the culture from which the image was drawn, the greater the accuracy in reproducing structure and especially detail.
Context refers to the social and physical setting in which cultural information is transmitted. The physical and social context of transmission can mediate or alter the content of what is being transmitted.
The context in which this information was transmitted greatly affected variability in how it was remembered and subsequently retransmitted (Barth, 1987, 1990).
It is possible that artifacts transmitted within ritual contexts had conservative rates of change.
Mode refers to the process by which individuals transmit and acquire information.
Cavalli-Sforza and Feldman (1981) and Boyd and Richerson (1985) have modeled, that different modes of transmission can have dramatic effects on the rate of evolution of cultural information. For example, many-to-one transmission tends to slow the rate of change relative to one-to-many transmission (MacDonald 1998, p. 230; Shennan 2002).
Conformist transmission is a many-to-one system but with a particular type of bias, where the ‘‘many’’ represents those individuals possessing the modal or average behavior.
Rarity (or pro-novelty) biased transmission
Prestige-biased transmission – certain prestigious individuals, rather than the masses, are assumed to have access to (or have experimented to acquire) superior information.
The mode of transmission can vary depending on how information is packaged.
In some cases, cultural information may be transmitted because it ‘‘hitchhikes’’ with other information (O’Brien and Lyman 2003).
Individuals (i.e., actors) receive information and intentionally act upon it (e.g., ignore it, choose from whom to accept it, modify it, experiment with it).
Example: Cultural transmission in material artifacts of archeology
Style is something that exists independent of an observer, that is, it is empirical.
Meltzer (1981, p. 314) suggests that ‘‘in many instances, the choice between certain kinds of design elements on ceramics is not a functional consideration, but rather is historically determined and selectively ‘neutral,’ because there is no inherent advantage between one element and the next. The actual presence of the design, however, has a selective value because that particular design serves to mark a certain individual or group boundary (or whatever other function it may serve).’ Style is, therefore, a way of measuring and explaining material culture through the conceptual framework of cultural transmission.
Fundamentally, a random copying model is a null hypothesis.
Mithen (1997, 1998): humans link the difﬁcult information to things they can easily transmit as part of packages.
Certain kinds of information will be bundled during transmission.
According to Mithen’s argument, the presence of such domain-speciﬁc structure within the brain results in strong patterns in covariation between certain kinds of information during cultural evolution.
Henrich (2004b) shows that the effective population size (i.e., the number of interacting social learners) is an important factor in the transmission of complex versus simple material technologies. Henrich ﬁnds that complex technologies tend to be lost when populations decrease in size while simple technologies are maintained or even improved.
Various transmission processes produce different patterns in variation, with some such as conformist transmission removing variants from the pool of behaviors (i.e., winnowing away), and others such as experimentation and innovation adding new ones.
The documentation of variability and measures of dispersion (e.g., standard deviation, coefﬁcient of variation) and covariation are not systematically reported in archaeological research. To maximize the utility of CT, it is important that archaeologists consistently report and consider the explanatory implications of dispersion measurements as well.
As a conceptual framework, CT is especially powerful for explaining patterns observed in material culture and variation therein through time and space.
Explanations of artifacts include the aggregate of ideas and processes involved in construction and how these are transmitted between individuals while simultaneously being modiﬁed through copying error, individual learning, experimentation, or innovation (e.g., Basalla 1988).
Three measures of variation are of particular relevance to CT.
The ﬁrst is the dispersion about a mode or average. Dispersion can be caused by a number of transmission processes such as purposeful experimentation or copying error.
The second measure of variation concerns the diversity of distinct types within an assemblage of artifacts.
The third measure of variation involves covariation between the attributes of an artifact or between artifact types themselves.
Few studies have explored the intersection of these three measures of variation (dispersion, diversity, and covariation).