What do we know about Cultural Transmission?

As I have been working on my book-in-progress (Breeders, Propagators, & Creators), I have encountered a difficult-to-answer question of road-block proportions: how do we quantify cultural transmission? The focus of my book is the tradeoff humans face between making babies, spreading existing culture, and inventing new ideas. If such a tradeoff exists, we need to be able to measure “success” in each of these endeavors. The biological part of success is relatively easy to measure: there are multiple ways to estimate the reproductive fitness of different people. Although this data is sometimes hard to track down (especially in the form that I need!), at least we know how to measure the breeding success of individuals. But what about the propagation of culture? While there’s a lot of scholarship on cultural transmission, a lot of that scholarship is about the very real difficulties associated with estimating cultural success.

So I was excited to be able to get my hands on a copy of Cultural Transmission (Schönpflug 2009a), an edited collection that explores a variety of approaches to understanding and investigating the process of cultural transmission. Because I have a pretty specific agenda regarding what I was looking to learn from this book, I did not read it cover to cover. As such, consider this a review of what this book contains that I am interested in rather than as a comprehensive book review.

How should we conceptualize cultural transmission?

The book provides a really crucial framework for conceptualizing culture, summarized in the An Ecological Perspective on Cultural Transmission chapter (Berry and Georgas 2009). According to this framework, culture is propagated from five sources:

1. Parents (vertical transmission);
2. Peers within one’s own culture (horizontal transmission, socialization);
3. Other Adults and Institutions within one’s own culture (oblique transmission, enculturation and socialization);
4. Peers from other cultures (horizontal transmission, acculturation); and
5. Adults and Institutions from other cultures (oblique transmission, acculturation).

There are several aspects of this framework that are different than the way that I have considered culture thus far.

First, the addition of oblique transmission to the categories of vertical and horizontal is interesting. Up until now, I have categorized all forms of cultural transmission as either vertical or horizontal. This over-simplification on my part probably stems from the way I think about culture: what matters is whether or not it is passed on to offspring. This means that I consider all forms of non-vertical transmission as “horizontal”, even though that non-parental culture can come from individuals either of the same generation or an older/younger generation. I can see where this source matters, as there has to be a net flow of culture from older to younger generations; whether that net flow is predominantly oblique or vertical in nature tells you a lot about the relationship between biological and cultural reproduction. An important issue for my analysis of the current state of human evolution is how decoupled cultural and genetic transmission are. If most cultural transmission is vertical in nature, that would mean that cultural and genetic transmission are fairly coupled. However, if cultural transmission is predominantly oblique, that suggests a decoupling of culture and biology… and therefore a potential conflict between the two systems governing human evolution.

A second aspect of this framework that I find interesting is the way that it delineates boundaries between cultures. We can reduce cultural transmission to only three sources if we do not bundle groups of ideas into particular “Cultures”; if we allow for discrete boundaries between cultures, we need special forms of oblique and horizontal transfer that emerge from outside of the recipient’s “home Culture”. Again, the novelty that this way of looking at culture presents to me betrays my own preconceptions of how culture should be conceptualized. Although I understand that talking about “Culture” with a big C is common in anthropology, I have always found this kind of categorization fairly arbitrary. If you consider individual people to be the reservoirs for cultural ideas (which I do), then any attempt to place these people into distinct cultures seems to me like a fool’s errand because belonging to an identifiable cultural group is no guarantee of being in possession of a particular set of cultural ideas.

I get that people who identify with a particular culture (for example, American) are more likely to have taken on particular cultural ideas than someone from a different culture (for example, Danish). But pull an American at random from our population and try to predict their cultural values (perhaps compared to the Dane) and the most you can hope for is to have a better-than-random chance of making a correct prediction. This is to say that large-scale cultures are not monolithic. At best they are the cultural analogue of regional human populations, who have a higher probability of sharing genes in common due to more recent common ancestry and/or geographical isolation. At worst, big-C cultures might represent something more akin to the cultural analog of human racial categories, arbitrary labels that turn out to have little biological meaning (Bamshad & Olson 2003). I guess that my fear is that trying to delineate anyone’s culture is impossible, especially now that there is so much flow of culture across geographical boundaries. It would not be surprising to learn that people who live in close proximity to each other are more likely to share cultural values, but that does not suggest that these similarities create emergent, coherent cultures that individuals can cleanly be labeled as “inside of” or “outside of”.

This is not to say that I think that only the individual scale of culture needs to be considered. Of course culture is a social phenomenon, and therefore can have emergent dynamics at the scale of various larger “groups”. Perhaps the most interesting group that Berry and Georgas bring to light is the “institution”. If there is any meaningful cluster of culture that supersedes the individual level, it is represented by social institutions that are well-defined by a particular set of cultural ideas. Religions certainly qualify, as they bundle an almost-dizzying array of ideas under “one big tent”. Governments of various scales and schools also serve to propagate a defined set of ideas. For this reason, understanding how institutions are culturally created and then how they serve to transmit cultural ideas and values is critical to understanding the nature of human culture. But again these institutions are not monolithic; people pick and choose which cultural ideas they choose to take on when they engage with an institution. There are plenty of Catholics on birth control and evolutionary biologists who believe in some kind of creative god.

How do we detect and quantify cultural transmission?

If you want to estimate the reproductive fitness of individuals in a population, you have a couple of options: witness or trace. You can witness individuals (usually females) giving birth and then keep track of how many of those offspring make it to a reproductive age. Or, you can trace overall genetic success by using DNA sequencing to determine (sometimes generations later) which individuals were most reproductively successful. Although neither method of estimating reproductive fitness is without its complications, it is important to note the very clear and discrete opportunities that our system of genetic evolution provide for assessing success.

Culture provides neither of these avenues, at least not with anything resembling the same clarity. Individuals pass on their genes in discrete moments and in large, discrete packages (called offspring!) whereas individuals pass on their cultures continuously and in difficult-to-quantify forms. While it is totally reasonable to posit that that cultural evolution is driven by differences in cultural fitness, measuring that fitness is a problem.

So how is cultural transmission measured in the chapters of Cultural Transmission? Although the details of this answer vary from chapter to chapter, overall the approach is one based on establishing correlations. How correlated are the values of parents and their children?; that is the most common question asked and answered empirically in this volume.

At the coarsest scale of resolution, Berry and Georgas (2009) analyze how the connectedness of families varies along several critical ecocultural variables (percentage working in agriculture, affluence, and access to education). They found strong correlations that suggest that in countries with a higher proportion of citizens working in agriculture, families maintain closer ties. In contrast, both affluence and educational access reduce the intensity of family ties. What does this have to do with cultural transmission? Well, if one assumes that closer family ties lead to higher-fidelity transmission of cultural ideas, understanding these broad trends might help us to predict the relative balance of vertical to horizontal/oblique transmission in different societies. The results are not particularly surprising, as we would expect more agricultural communities to both maintain a more local social orientation and to depend more heavily on the family as the social and economic unit. As individuals gain more access to education and become more affluent, their dependence on the family goes down. This trend is actually quite interesting because it suggests that type of cultural transmission may be correlated with overall reproductive success. Who has the highest reproductive rate on the planet overall? It’s people who live in the areas with the highest level of agricultural work and the least access to affluence and education. According to the findings of Berry and Georgas, these people are also most likely to translate their reproductive success into successful vertical transmission. These results need to be treated with caution, as they are so coarse that any number of confounding variables might be producing these correlations. As you will hear me lament over and over again, it would be so much more valuable to have data on both the reproductive success and intensity of family ties for a large dataset of individuals; such a dataset would replace coarse-scale correlations with individual-level correlations (perhaps!) and allow us to establish what (if any) trade-off exists between breeding and propagation.

At a slightly less coarse scale, Bernard Nuack’s chapter (2009) Intergenerational Transmission, Social Capital, and Interethnic Contact in Immigrant Families shows that the intensity of vertical cultural transmission can vary significantly across different immigrant groups who have emigrated to different countries. The immigrant experience creates an intriguing case study for understanding the nature of cultural transmission, as the strength of vertical transmission influences the balance between retention of cultural ideas from the country of origin and acculturation of values from the new country. Nuack considers the correlation of various culturally-mediated characteristics between two family pairs: fathers and sons and mothers and daughters. Although levels of correlation vary widely, some immigrant groups seem to maintain stronger vertical cultural transmission than others. There are relatively high correlations for some immigrant groups in measures such as educational degree, family-language retention, and feelings of discrimination, but these correlations do not hold for all immigrant groups. Perhaps the most interesting is how various measures of social connectivity correlate between generations. Network size and social capital (access to social connections via the social network) tend to be highly correlated across generations, which suggests that the potential network across which culture is transmitted is itself relatively heritable. As with the study by Berry and Georgas, Nuack’s study is far from considering the fidelity with which actual cultural values are transmitted vertically; all of the variables considered are more like cultural end-products than actual cultural variants.

A chapter entitled The Transmission Process: Mechanisms and Contexts by Ute Schönpflug and Ludwig Bilz (2009) illuminates some of the content biases associated with vertical transmission. Using regression analysis to detect concordance between father-son values in three populations of Turkish ancestry, Schönpflug and Bilz demonstrated that collectivist values tend to be vertically transmitted with far greater fidelity than individualistic values. This is interesting because it demonstrates that what values parents espouse influences the probability that those values will be embraced by their offspring. Parenting style was also shown to influence the probability of vertical transmission, as parents who were rated as less authoritarian and more empathetic were more successful at transmitting collectivist ideas to their offspring. These findings in total suggest that children in some way filter the cultural values espoused by their parents based on the internal consistency of these values with parental behaviors: it is hard to teach your kid to take on collectivist values when you as a parent in an authoritarian (in other words, highly hierarchical) manner. Schönpflug and Bilz refer to these kinds of transmission-biasing factors as “filters”, and it is interesting to think about how and why such cultural transmission filters may have evolved in development process of adolescents.

In their chapter (Accounting for Parent-Child Value Congruence: Theoretical Considerations and Empirical Evidence), Ariel Knafo and Shalom H. Schwartz (2009) point out another important modifier of vertical transmission: the accuracy with which children perceive the values of their parents. Unlike genes, culture is transmitted with notorious lack of fidelity. One source of poor fidelity transmission is misinterpretation of parental values by children. Knafo and Schwartz also sought to investigate the role of children’s acceptance of parental values in cultural transmission. In a detailed study that parsed the effectiveness of vertical transmission between both mothers and fathers and their sons and daughters, it was found that the degree of accuracy of perception of parental values and the degree of acceptance of parental values both substantially accounted for the degree of congruence between parental and children’s values. Not surprising results, but important nonetheless; as with the study by Schönpflug and Bilz, these results made me wonder about the factors that influence children’s perception and acceptance of parental values. Certainly we must have evolved to make prudent, context-specific decisions about when and when not to take on parental culture. Also echoing the findings of Schönpflug and Bilz, Knafo and Schwartz found that different values were perceived and accepted to varying extents; some values are much more likely to be accurately perceived and readily accepted than others.

Bernd Six, Kristina Geppart, and Ute Schönpflug (2009) also show that context matters in the transmission of particular kinds of cultural values. Their chapter, The Intergenerational Transmission of Xenophobia and Rightism in East Germany, considered the factors that influence transmission of xenophobic values from parent to offspring. Interestingly, parents were more likely to transmit their values to daughters than to sons, suggesting that horizontal or oblique tranmission may be stronger factors for boys than girls. Whether a child picks up xenophobic attitudes also depends on the other values he or she holds, again suggesting that human brains tend to favor the acceptance of novel values that are consistent with the existing suite of values previously accepted.

In their chapter Intergenerational Transmission of Violence, Haci-Halil Uslucan and Urs Fuhrer (2009) considered the degree to which parental violence directed at children influences the violence experienced and perpetrated by their children. Interestingly, while abused children were more likely to act violently towards others, they were even more likely to become the victims of violence outside of their homes. What is fascinating about this form of cultural transmission is that it is behavioral: we are not just talking about internal values but values embodied by actual behaviors. For cultural evolution to mean something, internalized values need to translate into consequential behaviors, and in the case of violence this translation seems to be occurring.

Tom F. M. ter Bogt, Wim W. J. Meeus, Quinten A. W. Raaijmakers, Frits van Wel, and Wilma A. M. Vollebergh (2009) consider the relative role of horizontal versus vertical transmission in their provocatively-titled chapter, “Don’t Trust Anyone over 25”: Youth Centrism, Intergenerational Transmission of Political Orientations, and Cultural Change. Apparently social scientists have identified the fact that adolescents vary in their degree of “youth centrism”, a measure of how much they identify with the culture of their peers. Ter Bogt and colleagues wanted to ascertain how much a tendency towards youth centrism influenced vertical transmission: are kids who orient towards youth culture more likely to obtain their values through horizontal transmission? Looking at two political values — tolerance of alternative lifestyles and equality of income and property — these investigators did find that higher youth centrism did lead to lower congruence with parental values, especially for tolerance of alternative lifestyles. However, even strong youth centrists showed an overall positive correlation of values with their parents, suggesting that youth centrism does not necessarily lead to rebellion against parental values. What I find interesting about this particular study is that it implicitly assumes that horizontal transmission may run counter to vertical transmission. While this is certainly possible, it is also possible — especially in a society with little variety in values — that what adolescents are likely to absorb from their parents is about the same as what they are likely to absorb from their peers. It’s only when a strong intergenerational difference in cultural values develops that the distinction between horizontal and vertical transmission becomes important.

In the Value Transmission and Zeitgeist Revisited chapter, Klaus Boehnke, Andreas Hadjar, and Dirk Baier (2009) consider whether the “modal current value climate of society” (how they define zeitgeist) influences cultural transmission. They use scores on a survey of Hierarchical Self Interest (HSI) as their cultural variable; HSI is roughly a measure of competitiveness, success orientation, and Machiavellism, so it has to be the product of many different cultural values that contribute to variations in HSI score. Does the most prevalent social value, the zeitgeist, influence value transmission? If considered across the whole of society, it does not appear to have much of an influence: HSI correlations were explained far better based on parental values than by the zeitgeist. But cleverly these investigators sought to look at the influence of zeitgeist in a slightly more nuanced manner: perhaps zeitgeist does not represent a predominant mode of cultural transmission but instead modifies the mode of vertical transmission within families. This sort of modifying effect was exactly what was found: in families whose HSI was close to the societal mode, vertical transmission of HSI was extremely weak, but in families whose HSI was far from the societal mode, vertical transmission was fairly strong. This is an intriguing result that has a lot to say about how cultural hegemony as well as cultural diversity are maintained. If your parental values are very much in sync with the predominant values of your society, as a parent you may not invest heavily in transmitting your values to your children: because your values are shared by the mainstream, your kids are almost as likely to gain the same values via oblique or horizontal transmission as they are via vertical transmission. But if your values are far from the mainstream and you want your children to share your values, you had better put effort towards vertically transmitting those values. These findings are fascinating and have important implications in a variety of social contexts. Recent immigrants, whose cultural values are quite likely to differ from those of their adopted homeland, are far more likely to expend effort propagating their cultural ideas to their children. Really anyone who is outside the mainstream will have to expend extra effort to see their ideas spread. Immediately I wonder if this means that the “propagation budgets” of different parents vary in relation to whether those parents share mainstream or minority cultural values; is it possible that parents within the mainstream spend more of their resources to propagate ideas horizontally and obliquely than parents outside the mainstream, who must focus more effort of vertical propagation?

As you can see, there are a lot of interesting, clever, and socially-meaningful studies of cultural transmission being conducted. They shed light on the patterns by which different kinds of cultural values are transmitted. But what do they tell us (if anything) about the general nature of cultural transmission?

Do we know anything about the predominant means by which culture is transmitted?

A key question for my Breeders, Propagators, & Creators enquiry is whether culture is mostly transmitted via vertical or non-vertical transmission. This question is important because — as implied above — how coupled biological reproduction is with cultural transmission tells us a lot about the nature of the tradeoff between breeding and propagating. If the best way to propagate your cultural ideas into the next generation is to have children, then our story is pretty simple: the winners of both the biological and cultural contests are those who produce the largest number of offspring. But there is good reason to believe that our biocultural story is not so simple: some of the most important forms of cultural transmission may be oblique or horizontal, implying a decoupling of — and thus tradeoff between — cultural and biological (re)production.

What light do the empirical chapters of this book shed on this question? Well, these chapters mostly show that there’s a lot of nuance to vertical transmission. To think simplistically about vertical transmission as making a particular contribution to the overall transmission of culture across generations is to miss out on a lot of contextual variation. How much vertical transmission matters can depend on the economic activity of a family, the location of its cultural values relative to those of the larger society, parents’ and childrens’ gender, the generational orientation of children, existing values held by children, and the actual content of the values being transmitted. Not all values will be vertically transmitted with the same rate of success in different cultural environments. Understanding this cultural context-specificity is important.

But while many of these studies explore subtle nuances of vertical transmission, none directly gets at the question of how much vertical transmission matters in comparison to oblique or horizontal transmission. The closest that any of these studies gets to directly looking at vertical versus non-vertical transmission is the study on “youth centrism” that demonstrates that a social orientation towards one’s peers translates to lowered vertical transmission of values from one’s parents. Presumably this means that the values of “youth centric” adolescents were more likely to be received horizontally, but such “conclusion from presumption” is underwhelming. For many of the cultural values explored in these studies, parent-offspring correlations were relatively low. This suggests low vertical heritability of cultural values, and again leads to the very soft inference that other forms of transmission must dominate. But because all of these studies focus on parent-offspring correlation of cultural values, we have no direct measure of the influence of peers, or non-parental adults, or social institutions.

A lot of these studies look at the cultural transmission of very complex values or value systems. For example, one’s score on the survey of Hierarchical Self Interest (HSI) cannot emerge solely from carrying a few cultural values. Instead, one’s HSI score reflects variation in a variety of different cultural ideas in a variety of potential cultural dimensions. As such, looking at the transmission of such cultural complexes is a bit like trying to decipher the heritability of a polygenic trait: just measuring phenotype doesn’t tell you a whole lot about the role of underlying genetic variation in influencing that phenotype. Understanding how complex polygenic traits evolve is a substantial challenge for evolutionary biology; similarly, we should expect it to be a lot more difficult to understand the nature of cultural transmission if we are constantly considering cultural complexes. I am well aware that the meme does not exist as a convenient analog to the gene, but I do wonder if we could learn more about the nature of cultural transmission by focusing on more atomized cultural values rather than large, somewhat-nebulous concepts like “xenophobia” or “tolerance of alternative lifestyles”. Although these big-culture concepts are perhaps more important to understand from a sociological perspective, if you want to understand the underlying mechanisms of cultural evolution it is probably necessary to track smaller units of cultural inheritance.

I am not suggesting that the methods used in these collected studies are invalid (although some of them produce very tepid conclusions). Instead, what I am recognizing is that these sorts of cultural transmission studies don’t address the kind of questions about cultural propagation that I am interested in.

What I am still looking for…

What I was hoping to gain from reading Cultural Transmission was an understanding of how cultural researchers track the process — and not just the results — of cultural transmission. After reading through this collection, I came to see that a more accurate title for this book would have been Transmitted Culture. Because of the very high reliance on correlation studies, most of what we learn about is the product of cultural transmission rather than transmission per se. And if we really want to understand the nature of cultural fitness, we have a problem. When we see high correlation between parents and offspring, how do we explain this correlation? All we can say is that it is possible that parents have successfully transmitted their values vertically. But it is also impossible to distinguish this potential vertical transmission from horizontal or oblique origins, especially if a complex cultural measure is common in a society.

What I really want to see are studies that track the transmission of culture in a manner that’s analogous to how we can track genetic fitness. Because culture leaves nothing like the distinctive signature of genetic inheritance, the “trace” option is not available for culture. To really understand the nature of cultural transmission, we need the ability to witness rather than trace. And witnessing the direct transmission of culture is not going to be possible very often. Perhaps the most exciting opportunities for bearing direct witness to the cultural transmission process come at the genesis of a new cultural idea. New words, new ideas, and new tools all provide us with the opportunity to witness cultural transmission as it happens because — unlike most cultural variants — new ideas are not so widespread as to be impossible to follow. An added benefit of following the spread of novel variants is that doing so would shed light on the fate of newly-created culture, helping us to understand how often “cultural mutations” manage to survive.