Does God’s omnipotence extend to vision?

In a recent paper, Lorenza Colzato and her colleagues have tried to provide some pieces of evidence for cultural influences on perception, and more precisely, for their hypothesis that following religious rules can 'affect' visual processing.

I am very sceptical about this paper. One crucial question I have been obsessed with while reading it was: how can high-level things like beliefs be influential at such a low level as visual perception? For sure, Colzato reports evidence for supporting this hypothesis (as did a previous paper of Colzato's) but what remains unclear is the function such an influence could fulfil, and how this influence could be cognitively realized. Moreover, the authors report that this influence of beliefs on visual processing is not specific to religion.

KuyperArguing that "cultural contexts are very hard to capture and to define" (since citizens of a same country could belong to different cultures, a geographical cut-out is not completely sufficient to distinguish between cultural areas), Colzato et al. have decided to focus on religion, and they decided to define a 'religion' as a set of rules. Thus, every follower of a religion is tied to some rules which are, little by little, inducing "particular cognitive-control strategies and establishing default control parameters that generalize to situations that have no bearing for religious beliefs". This argument is supported by other studies such as McCullough and Willoughby's, in 2009, which shows that religious people were less likely to break the law since they were more 'used to' following rules in general (because of their daily, effortful religious training).

Yet, the authors argue, these induced cognitive-control strategies are not limited to complex decisions (such as "to break or not to break the law"). They can be found in low-level processing mechanisms like vision.

For example, Colzato et al.' previous study (already linked to) has shown that Calvinists (whose cultural background is supposed to stress the notion that every sphere of the society should mind its own business) were more likely to be victims of the global precedence effect than were atheists (who, the authors suppose, possess a more "holistic" view of the society). In the global precedence effect, as described by D. Navon in 1977, the global structure of a percept is available earlier than its local features, for example, to see the forest before the trees.

(Photograph: Abraham Kuyper, father of the Dutch Anti-Revolutionary Party, a conservative Calvinist movement.)



In this new paper, the aim of Colzato et al. is to demonstrate that this effect of religious precepts is long-lasting, and that it crucially depends on how strict the religious training has been (a second study – dedicated to show that this effect was not a 'religious effect' but specific to one religion in particular, Calvinism -, was proposed. Here, I am focusing on the first study).



To test the 'chronic influence' of religious training, the authors compared the 'global precedence effect' in conservative neo-Calvinists, liberal Calvinists, baptized atheists (in Calvinism), and non-baptized atheists. The prediction was that this effect would be less important in Calvinists than in non-baptized atheists. It would be more important in non-baptized atheists than in baptized ones. Because Calvinists are supposed to have a daily practice of the rules of their religion (and above all, according to the authors, the principle of individual responsibility in every sphere of life), they will focus more on the details than baptized and non-baptized atheists do. In the same manner, baptized atheists, because they received a Calvinist education when they were younger, should be affected by this bias, but to a lesser extent.

The second purpose of this first study was to show that the strictness of religious devotion was negatively correlated with the bias: the stricter the Calvinists were (subjects were divided between "liberal" and "conservative" Calvinists), the less important the bias would be. 
The results confirmed all the predictions – except for the fact that baptized atheists were not significantly more likely to be victims of the global precedence effect than liberal and conservative Calvinists, suggesting that an early exposure to a religious environment could durably affect the visual processing mechanisms. For the authors, these results provide evidence for a robust bias on perceptual processing, due to the commitment to a religious way of life, and depending on the strength of this commitment.

Several criticisms can be made of that conclusion.

First, the claim about individual responsibility is probably not exclusively a Calvinist rule. Colzato mentioned the "pillarisation" (" verzuiling" in dutch) of the Dutch society, that is to say, the fact the Dutch society is organized on a community-by-community basis. Each religion (or ideology) has their own newspaper, banks, schools, hospitals etc. In the Netherlands (where Colzato's lab is settled), the society is organized between three main pillars (Catholic, Protestant and Socio-Democrat). Even if the pillarisation of the society is an idea which was above all promoted by the Protestants of the Anti-Revolutionary Party ("Anti-Revolutionaire Partij", whose father Abraham Kuyper was a fervent advocate of the idea that a multiple religious country –like the Netherlands- should be segregated into different societal institutions), this thesis is now shared by many non-Calvinists. Since the life of most individuals in Netherlands is organized according to the principle of pillarisation, we should not expect, on that basis, a difference between Calvinists and non-Calvinists regarding to the global precedence effect.

The "individual responsibility of each sphere" doesn't necessarily entail the responsibility of every individual within communities. We could easily imagine that a true liberal person (who believes that every individual in the society –and not every sphere of life- should mind her/his own business) would be even less victim of the global precedence effect.

Finally, Colzato's conception of 'atheism' can be criticized. Atheism is not a set of rules, neither is it a doctrine. There is no reason why every atheist should rely on 'collective action', should support 'solidarity', and should promote 'holistic points of view'. Therefore, the difference in visual processing could be explained by a greater exposure to a particular environment. This 'exposure effect' could explain the difference between baptized atheists and non-baptized atheists too, as well as the difference between conservative and liberal Calvinists.

One last critic could be made about the concept of 'religion' used here. Unfortunately, focusing on 'religion' is not the best strategy one can adopt to make the question of cultural influence on visual processing easier: one controversial point is whether a religious phenomenon can merely be defined as a precise set of rules, and whether these rules are effectively and precisely embodied into practices. For example, one could define religion as a set of beliefs, assuming that some of these beliefs will never be incarnated into concrete and material practices (such as "God is omniscient"). How do these beliefs could have an effect on visual processing if none of them has a real impact on the daily life on their followers?

In any case, this paper raises several questions: can cultural life shape such low-level processing (a long-standing debated question)? If it is the case, how could this cognitively work? Aren't there parts of bits or parts of cultural life which are more influential than the others? At least, what function could be attributed to these culturally shaped visual bias?


Colzato et al.'s paper:

Colzato, L.S., van Beest, I., van den Wildenberg, W.P.M., Scorolli, C., Dorchin, S., Meiran, N., Borghi, A.M., Hommel, B. (2010). God: do I have your attention? Cognition, Volume 117, Issue 1, October 2010, Pages 87-94 (link)

Their previous one :

Colzato, L. S., van den Wildenberg, W., & Hommel, B. (2008). Losing the big picture: How religion controls visual attention. PLoS ONE 3(11). (link)

Religion and rule-following

McCullough, M. E., & Willoughby, B. L. B. (2009). Religion, selfcontrol, and self-regulation: Associations, explanations, and implications. Psychological Bulletin, 135, 69–93. (link)

The global precedence effect

Navon, D. (1977). Forest before trees: The precedence of global features in visual perception. Cognitive Psychology, 9, 353–383. (link)


  • comment-avatar
    Tom Rees 18 October 2010 (10:08)

    Colzato also found that devout Catholics and practicing Budddhists had more global visual attention than their less-religious countrymen. So it’s not so much that “religion” has a specific effect, but rather that the cultural setting (as reflected by religion) has an effect. What, precisely, that cultural setting is is open to question, as you say!

  • comment-avatar
    Simon Barthelme 24 October 2010 (16:26)

    I agree with Guillaume’s points – assuming the phenomenon is real there is a huge gap in the story. Either religious practice influences the environment of a person enough to warrant having different attentional “default settings”, and I’d be verious curious as to what those changes in the environment could be, or religious ideas somehow affect the visual system directly (in a purely idealistic process, so to say). The latter possibility seems to me unlikely. From the experimental point of view, the paper is missing a control – what they describe as local/global differences could simply be due to some groups being more motivated than others. Since the global task is really easy (look at the error rates), this differences in motivation would show up only in the – harder – local task. In other words the experiment might just suffer from what is called a “floor effect” in the jargon. They need an extra task, say tone detection for example, on which they show that no performance differences as a function of religious practice is expected and none is found.

  • comment-avatar
    Guillaume Dezecache 2 November 2010 (11:20)

    Thanks for you comments Tom and Simon. Tom: I am not more convinced by Colzato’s study on Buddhism. What lacks the most, I guess, is a precise account on the causal role of abstract beliefs on visual processing. As for your review of this paper, could you tell me what you precisely hear by “a selection bias”: “It might be that those people whose attentional bias doesn’t mesh well with the predominant religion are more likely to become atheists. That’s possible, although the fact that atheist converts differed from those raised as atheists in the Dutch study would tend to suggest that isn’t the case”. Do you mean that a visual bias could influence the way a religion will develop in a given society (if the population is affected by this visual bias) ? Simon: I totally agree with your comment, and especially on the lack of a control experiment…