Societal Value Culture

The four insightful commentaries raise many questions about my “think-piece.” They also use it as a platform to develop their own fruitful ideas about culture. I am delighted to have stimulated this discussion.

My own effort had two basic aims. First, I sought to explain why it is legitimate to characterize countries by their prevailing cultural values, despite the finding that the values of individuals vary more within- than between countries. Second, I sought to justify aggregating the values of individual societal members as a meaningful way to estimate the prevailing cultural values, despite low within-country consensus on individual values.

The key to understanding my position is to grasp the distinction I make between individual and cultural values. Like others, I view individual values as a psychological property of persons, located in their minds. In contrast, I view cultural values as a hypothetical, latent feature of societies (or groups), a normative system that is independent of individuals and is not located in their minds. We can infer cultural values from their manifestations in the complex of meanings, beliefs, practices, symbols, norms, and values prevalent in a society and its institutions. The Fischer and Schwartz (2011) findings refer to variation of individual values within- and between countries, not to cultural values.

The three commentaries that refer to the idea of aggregating individual value priorities and norms accept that this can provide meaningful information about culture, even in the absence of substantial consensus. Moreover, they too distinguish culture from properties of the individual, locating it in the social and symbolic context that influences and constrains individuals and to which they actively adapt. Morris rejects the idea of measuring culture as a latent variable, but Wan and Lu and Yamagishi do not find fault with this idea. Our overall agreement on basic premises is gratifying.

Many ideas and examples in the commentaries have enriched our understanding of culture. Below, I briefly comment on only a few of the ideas and criticisms.

A recurrent theme in all of the commentaries, especially in Wan and Lu, is the importance of considering the norms, values, and expectations people perceive others to hold as aspects of culture, in addition to individuals’ own values. They note that there is likely to be greater consensus regarding perceived norms and values than regarding individuals’ own norms and values. Moreover, these perceptions, even if mistaken, may be as influential as or even more so than the value culture reflected in aggregated personal values. Thus, the power of culture may depend more on the perception that people share beliefs and values than on actual sharing. I agree with all of these points and with the call for expanding the study of perceptions and for identifying the complementary roles that cultural values and perceived cultural values and norms play in societies and other groups. Moreover, aggregating the perceived values of societal members would provide an additional index of the latent societal value culture useful for cultural comparisons.

Another recurrent theme is the importance of examining the idea of cultural sharedness more closely. In reacting to the challenge of the Fischer and Schwartz (2011) findings, I was concerned with the extent to which individuals share their value priorities. Wan and Lu raise the interesting question of the adaptive function of shared individual values. Especially promising for future research is their idea that the extent to which specific values are shared may depend on their relevance to the adaptive challenges that particular societies or institutions face.

A third recurrent theme that I strongly endorse is the existence of multiple, proximal cultures (e.g., ethnic, professional, religious, family) that influence individuals more strongly and directly than any overarching value culture. The model I developed in the figure in my article includes the level of institutions for just this reason. Although influenced by the prevailing societal value culture, every institution, large and small, develops its unique value culture as a function of its specialized tasks, challenges, and membership. Because individuals are exposed to and/or choose to locate themselves in different sets of institutions, each person adapts to a different proximal cultural press. Greenfield makes a similar point by suggesting that the presence of groups with different sets of sociodemographic characteristics accounts for much variation in values within societies.

Yamagishi criticizes my choice of the measurement of latent aptitudes as an analogy for deriving cultural value scores from individual value responses. He suggests that it breaks down because of measurement nonequivalence and because it is too static. I do not understand the first critique. The value items used in my cultural value work have demonstrated equivalence of meaning across cultures. As long as the samples of respondents accurately represent their groups (or countries), the analogy of aptitude items seems to me to work well. But no matter, it is only an analogy. It is not the theoretical model depicted in the figure. If Yamagishi’s alternative analogy persuades those who find aggregating individual responses problematic, fine. The measurement analogy is indeed static, but the theoretical model is not, as I elaborate below.

Greenfield’s “cultural values” are what I call individual values. She attributes the greater variation of values within countries than between countries to the “invariably . . . great[er] socio-demographic variability within their borders.” Adaptation to sociodemographic variability certainly accounts for some of the within-country differences in values. The downward-pointing arrows in my figure signify this idea. But these forces may not affect individuals’ value priorities more strongly than their unique personal characteristics, their personality, temperament, genes, health, and so on. Equally important, the particular ways that individuals adapt their values to their environment depend on their personal characteristics.

In her persuasive account of historical trends toward individualism, Greenfield suggests that the global growth of individuation has attenuated value consensus within societies. Thus, it is probable that the ratio of within-country to between-country variation in values has increased. She further suggests that data from earlier decades than Fischer and Schwartz (2011) studied might not have shown greater variation of values within- than between countries. Undoubtedly, individualistic values have become more important at the expense of collectivistic values in recent decades. Whether this shift has brought about an increase in the within-country variation in values is an empirical question. The seven value items from the World Value Survey (WVS) used by Fischer and Schwartz permit an initial test.

For each of these values, I computed the within- and between-country variance in the 14 countries that participated in the 2005-2007 WVS wave and in either the 1981-1984 or 1989-1991 wave. Many of these countries experienced major social change during these years (e.g., South Africa, South Korea, China, Russia, Mexico). The minimum period between the two times was 15 years and the mean was 22 years. On average, individual differences explained 80% of the variation in values at the earlier period versus 73% in 2005-2007. Thus, earlier and later data than the 1995-1996 data used by Fischer and Schwartz (2011) lead to the same conclusion that challenges the “shared values” perspective. For five of the seven items, within-country consensus increased over time. Across all items, the ratio of within-country to between-country variation in values dropped from 6.7 to 4.9. In sum, despite the likely rise in individuation, value consensus within countries increased rather than decreased over two decades.

Explaining why Greenfield’s plausible speculation is not supported requires further theorizing and research. A study of the effects of socioeconomic development and democratization on within-country value consensus (Schwartz & Sagie, 2000) gives some hints. Across 42 countries, they found less overall value consensus as a function of democratization but more value consensus as a function of socioeconomic development. These two structural variables tended to neutralize one another, but their effects on specific values were not all the same. This suggests that it is crucial to focus on specific values when examining the effects of socioeconomic and historical change on value consensus.

In his rich and wide-ranging essay, Morris offers several criticisms of my article that require a response. Some of these arise because he attributes to me positions that I do not hold. I accept responsibility for this as reflecting a lack of precision on my part. Had I written the word value in front of the word culture every time I used the latter, we might have avoided many misunderstandings. Thus, contrary to what he attributes to me, I do not define culture narrowly as limited to values. Furthermore, I treat the absence of consensus on values as a challenge only to views of culture in which shared meaning systems refer to agreement among individuals regarding the importance of various values to them (e.g., Hofstede, 2000; Inglehart & Baker, 2000). Of course, people within societies and institutions share the meaning of a multitude of concepts, making communication possible. I certainly do not “reduce culture to country differences.” Values are a key aspect of culture that has proven fruitful for cultural and cross-cultural psychology, but I do not claim they are “the essence of culture.” I am pleased that the other commentaries recognized that my emphasis on cultural values did not exclude additional aspects of culture that they discuss.

Morris worries that postulating a latent society-level value system may be imputing an essence to society, a ghost in the societal machine. He proposes instead that culture exists in the distribution of ideas, norms, values, expectations, artifacts, and so on., in a society. I see these as the observable manifestations of culture. The inferred value culture is hardly a ghost. Like any latent variable (e.g., personality trait), it is measured by its effects expressed in these manifestations, particularly, but not necessarily exclusively, in the values of individuals. Any factors that impact on a society—natural disasters, wars, technological breakthroughs, social inventions—can lead to measurable changes in the latent value structure. Latent variables are widely used in science to conceptualize variables that influence observable phenomena. They facilitate scientific advances by extracting crucial, theoretically meaningful elements from the array of observables.

Perhaps our difference regarding the use of latent variables reflects the goals of our research. I wish to compare the cultures of societies or other groups in order to understand differences between their policies, traditions, and practices and the attitudes and behaviors of their members. This requires broad cultural dimensions on which meaningful comparisons can be made. The dimensions I propose are value emphases that emerge as groups cope with the challenges of their social and material environment and with regulating the behavior of their members. I infer the locations of societies on these latent value dimensions by observing the effects of cultural values in individuals’ values. Morris is less interested in broad cultural dimensions that are useful for comparison. He seeks, instead, to understand the ways that various aspects of culture affect and are affected by the acts and ideas of individuals and groups as they communicate and coordinate in concrete contexts.

Based on the kind of careful reading that writers are rarely privileged to receive, Morris notes that, unlike my earlier work (Schwartz, 2006), my current model leaves out the possibility that cultural values might mediate effects of institutions on individuals. I thank him for noting this because it is an unintended narrowing of my views. In my attempt to present a simple graphic model, I placed the “latent culture” boxes at only one level in the figure. They should appear above and below the level of institutions. This would signify that, as my past work has shown, cultural values mediate many of the effects of institutions on individuals (e.g., of economic or political systems on the personal importance of religion and on acceptance of foreign workers, abortion, and divorce).

Finally, let me comment on the perception of my model as static (Yamagishi, Greenfield) or as underplaying conflict and change in societies because of its functionalist emphasis (Morris). My approach to indentifying key cultural value dimensions is indeed functionalist. I derived value dimensions for comparing cultures by considering three of the critical problems that confront all societies. These dimensions capture the value preferences or emphases that evolve as the people in societies plan responses to these problems and seek to motivate one another to cope with them. As my research across countries has shown, there is substantial coherence in the organization of value emphases in most countries (Schwartz, 2006). Where cultural egalitarianism is emphasized, for example, cultural autonomy tends to be emphasized too (e.g., in Western Europe), and cultural hierarchy and embeddedness are not encouraged. But tension, conflict, and change are inherent in my model. Individuals, subgroups, and institutions with difference interests and traditions are likely to promote behaviors, ideas, and aspirations incompatible with prevailing value emphases. The lighter arrows pointing upward at every level in my figure signify these sources of change.

Even more important for the dynamics of conflict and change are the impacts of exogenous factors on the cultural value system. My model of societal value culture describes an open, not a closed, functional system. Societies continually adapt to political, economic, military, climatic, technological, and other changes in their environment. In this process, their preferred value emphases change. My theoretical framework stresses this point no less than Greenfield’s. I too have studied the impacts of political and social shifts on cultural values (e.g., Schwartz, 2009; Siegel, Licht, & Schwartz, 2011). The prominently placed “Ecology, History” box at the top of my figure represents the exogenous impacts in this open systems model. The heavy downward-pointing arrows from the box to all other parts of the model emphasize that nothing is static. The weaker upward pointing arrows indicate that individuals, institutions, and cultures can reciprocally influence exogenous forces. Though broad cultural values usually change slowly, change can occur suddenly and rapidly in the face of major environmental challenges.

The four commentaries have helped me to sharpen my own thinking and to articulate points that were insufficiently clear. I hope that this exchange of ideas will stimulate others to rethink their understanding of societal value culture and of culture more generally.