As behavioral scientists, we have to come to terms with the fact that, if we want to survive in our culture, publishing and building a reputation that brings recognition, not to mention salary and grant money, it is better to account for what is stable and predictable. It often has better academic pay-off than trying to account for what is diverse and variable.
In academia, a priori claims of universality sell better than diversity, which complicates rather than simplifies matters. Universality claims get more attention because they are cleaner and sharper, projecting more encompassing control and predictive power in the field. Such claims are also better didactic tools. They have all-around greater impact and appeal. This tends to relegate diversity to noise rather than a primary object of study.
There is no clear escape from such reality. Yet, as scientists, we have to be constantly reminded of how bully-ish and presumptuous we intrinsically are in our reductions and generalizations. This is what Henrich et al. do pointedly in their target article. They deserve to be congratulated for their effort, in the footsteps of Arnett (Reference Arnett2008) and Rozin (Reference Rozin2009). However, there are some important points that they either omit, or at least seem to gloss over too quickly.
First of all, the serious issue of universal validity cannot be resolved by just testing a larger, more culturally diverse population. The sampling of diversity is crucial. The idea of setting up labs at international airport terminals, as recommended in the article, will simply not do it. Think about it: Who is traveling to foreign countries by plane? Those internationally “rich” ones would arguably be no more diverse (hence representative) than college students. They might be as “weird” and non-representative, not a good measure of what is universal or variable about the mind. As a measure of the sampling issue problem, be reminded that 80% of the world's population lives on a family income of less than $6,000 a year, with half of the world's population living on an average of 2 dollars a day (Kent & Haub Reference Kent and Haub2005; UNDP 2006). Who would claim that poverty has no impact on shaping the mind? However, dividing the world into rich and poor would include major confounds, and the question is where to set the limits of the population category boundaries?
The question, somehow eluded in Heinrich et al., is: how to sample diversity and what rationale to adopt? What are the criteria? Geographic? Ecological? Physiological? Socioeconomic? Demographic? Cultural? These are important considerations and, in the end, it all depends on the theory behind the quest for universality. Diversity cannot be controlled just for the sake of it, particularly if the ultimate motive is to determine what is invariant and what changes across populations. The delicate question, somehow glossed over by the authors, is: On what basis do we cluster populations? Such decision can often be “shaky.” For example, catchy acronyms like “WEIRD” for a population sample are good mnemonics. However, they carry the danger of distracting us from deeper issues. The last letter, D, for example, stands for “Democratic.” What does this mean, given that many Eastern cultures would not consider themselves as non-democratic, having universally elected parliaments in their countries? In using such an acronym to characterize a population sample, the authors must have a theory about what democrats and a democracy mean. They must also have some intuition as to what kind of impact such a regime might have on its citizens, as opposed to another. The democratic criterion would deserve more articulated rationale. It is not as easily measurable compared to race, wealth, or education, yet it probably captures something important. The question is what?
The name of the game in science is to generalize from limited population samples: samples that are easily accessible, easy to work with, and closer to the culture we as scientists know best. In this respect, there is nothing wrong about ethnocentrism. From such sampling we keep building paradigms on how “children,” “folks,” and “the mind” work. In the scientific context, the claim of universality is neither wrong nor presumptuous. It is a goal and a measure of scientific truth. The problem is that such a measure is always relative to its instrument.
What seems to be really wrong with a priori claims of universality is not the lack of population diversity, per se, as implied in Heinrich et al. It is the use of Western “WEIRD” instruments to measure behavioral phenomena across cultures. The use of Westernized procedures and experimental paradigms thought out to fit Western intellectual ecology and preoccupations (e.g., two-dimensional symbolic optical illusions such as the Müller-Lyer one, instead of ecologically more valid, hence potentially more universal, optic phenomena such as the “moon illusion”; the use of “weird” anonymous, and on the whole abstract, one-shot economic games; “weird” abstract questionnaires and other I.Q. testing) might well account for why typical Western samples are systematic outliers displaying extreme, non-representative behaviors. Behavioral sciences, historically a WEIRD practice, create culturally specific objects and tools of investigation to measure our predominantly WEIRD population against other populations. This is the true guilty aspect of ethnocentrism.
Last but not least, it is essential to distinguish levels at which universality claims may apply. The macro levels of economic games or self-assessment surveys are incommensurably more subject to population variability compared to lower level processes such as face recognition or emotional expression. This needs to be taken into consideration, the problem being not only how different populations are, but instead and more importantly, how populations vary in their distribution. But this distribution certainly depends on whether we are dealing with a low-level processing, such as emotion recognition, or higher level, such as self-concept.
To conclude, it would be disingenuous to think that we are not in the business of trying to approximate natural laws, whether these laws account for the homogeneity or diversity (context and culture dependence) of phenomena that, by consensus, need to be falsifiable. This is the context in which scientific claims of universality should be read and discussed. Whether population variability is ignored or not, the assumption of universality is true until proven wrong. As in any scientific debate, there is no end in sight.
You have
Access
As behavioral scientists, we have to come to terms with the fact that, if we want to survive in our culture, publishing and building a reputation that brings recognition, not to mention salary and grant money, it is better to account for what is stable and predictable. It often has better academic pay-off than trying to account for what is diverse and variable.
In academia, a priori claims of universality sell better than diversity, which complicates rather than simplifies matters. Universality claims get more attention because they are cleaner and sharper, projecting more encompassing control and predictive power in the field. Such claims are also better didactic tools. They have all-around greater impact and appeal. This tends to relegate diversity to noise rather than a primary object of study.
There is no clear escape from such reality. Yet, as scientists, we have to be constantly reminded of how bully-ish and presumptuous we intrinsically are in our reductions and generalizations. This is what Henrich et al. do pointedly in their target article. They deserve to be congratulated for their effort, in the footsteps of Arnett (Reference Arnett2008) and Rozin (Reference Rozin2009). However, there are some important points that they either omit, or at least seem to gloss over too quickly.
First of all, the serious issue of universal validity cannot be resolved by just testing a larger, more culturally diverse population. The sampling of diversity is crucial. The idea of setting up labs at international airport terminals, as recommended in the article, will simply not do it. Think about it: Who is traveling to foreign countries by plane? Those internationally “rich” ones would arguably be no more diverse (hence representative) than college students. They might be as “weird” and non-representative, not a good measure of what is universal or variable about the mind. As a measure of the sampling issue problem, be reminded that 80% of the world's population lives on a family income of less than $6,000 a year, with half of the world's population living on an average of 2 dollars a day (Kent & Haub Reference Kent and Haub2005; UNDP 2006). Who would claim that poverty has no impact on shaping the mind? However, dividing the world into rich and poor would include major confounds, and the question is where to set the limits of the population category boundaries?
The question, somehow eluded in Heinrich et al., is: how to sample diversity and what rationale to adopt? What are the criteria? Geographic? Ecological? Physiological? Socioeconomic? Demographic? Cultural? These are important considerations and, in the end, it all depends on the theory behind the quest for universality. Diversity cannot be controlled just for the sake of it, particularly if the ultimate motive is to determine what is invariant and what changes across populations. The delicate question, somehow glossed over by the authors, is: On what basis do we cluster populations? Such decision can often be “shaky.” For example, catchy acronyms like “WEIRD” for a population sample are good mnemonics. However, they carry the danger of distracting us from deeper issues. The last letter, D, for example, stands for “Democratic.” What does this mean, given that many Eastern cultures would not consider themselves as non-democratic, having universally elected parliaments in their countries? In using such an acronym to characterize a population sample, the authors must have a theory about what democrats and a democracy mean. They must also have some intuition as to what kind of impact such a regime might have on its citizens, as opposed to another. The democratic criterion would deserve more articulated rationale. It is not as easily measurable compared to race, wealth, or education, yet it probably captures something important. The question is what?
The name of the game in science is to generalize from limited population samples: samples that are easily accessible, easy to work with, and closer to the culture we as scientists know best. In this respect, there is nothing wrong about ethnocentrism. From such sampling we keep building paradigms on how “children,” “folks,” and “the mind” work. In the scientific context, the claim of universality is neither wrong nor presumptuous. It is a goal and a measure of scientific truth. The problem is that such a measure is always relative to its instrument.
What seems to be really wrong with a priori claims of universality is not the lack of population diversity, per se, as implied in Heinrich et al. It is the use of Western “WEIRD” instruments to measure behavioral phenomena across cultures. The use of Westernized procedures and experimental paradigms thought out to fit Western intellectual ecology and preoccupations (e.g., two-dimensional symbolic optical illusions such as the Müller-Lyer one, instead of ecologically more valid, hence potentially more universal, optic phenomena such as the “moon illusion”; the use of “weird” anonymous, and on the whole abstract, one-shot economic games; “weird” abstract questionnaires and other I.Q. testing) might well account for why typical Western samples are systematic outliers displaying extreme, non-representative behaviors. Behavioral sciences, historically a WEIRD practice, create culturally specific objects and tools of investigation to measure our predominantly WEIRD population against other populations. This is the true guilty aspect of ethnocentrism.
Last but not least, it is essential to distinguish levels at which universality claims may apply. The macro levels of economic games or self-assessment surveys are incommensurably more subject to population variability compared to lower level processes such as face recognition or emotional expression. This needs to be taken into consideration, the problem being not only how different populations are, but instead and more importantly, how populations vary in their distribution. But this distribution certainly depends on whether we are dealing with a low-level processing, such as emotion recognition, or higher level, such as self-concept.
To conclude, it would be disingenuous to think that we are not in the business of trying to approximate natural laws, whether these laws account for the homogeneity or diversity (context and culture dependence) of phenomena that, by consensus, need to be falsifiable. This is the context in which scientific claims of universality should be read and discussed. Whether population variability is ignored or not, the assumption of universality is true until proven wrong. As in any scientific debate, there is no end in sight.