Understanding causation in social/cultural contexts
#1
Often in conversations, when its pointed out that so and so is caused by culture or some social systems, some typical objections come up. I would like to dedicate this thread to counter these standard objections.

Before those counter arguments can be presented some definitions need to be understood
1. Necessary condition: A cause that is necessary for an effect. The effect cannot happen in absence of the necessary condition. E.g. wind is a necessary condition for wind mill's operation.
2. Sufficient condition: A cause that is sufficient for an effect. Given the cause the effect will certainly happen. e.g. Deadly venom is sufficient to cause death.

A point to be noted is that Necessary condition needn't be a sufficient condition and vice versa. i.e. Wind might be necessary for wind mill's operation but it is not sufficient. Proper functioning of other mechanisms is needed as well. Similarly deadly venom is sufficient to cause death but not necessary one. Death can be caused by various other reasons.

Coming back to the standard objections
Objection 1: Citing X person/group as exception
Objection 2: Effects exists even without the claimed causes.
Objection 3: Changing the claimed cause won't bring about change in the effect.

Lets consider an example that doesn't carry any moral baggage for easy evaluation.
Claim: Speakers produce music.
Objection 1: But I have a pair of 20 year old speakers that do not produce music.
Objection 2: But guitars do not have speakers and they still produce music.
Objection 3: Removing the speakers is not going to stop music as the live orchestra will still be audible.

The above objections sound inane because everybody understands what exactly the claim is and how the objections are not applicable in the correct context.

But now consider topics that get hotly debated.

1. Caste based discrimination leads to restricted growth opportunities
Objection 1: But Dr Ambedkar accomplished a lot even when there was a lot of caste based discrimination.

2. Misogyny in Indian culture leads to rapes
Objection 2: But this happens even in west.

3. Slurs lead to bigoted culture.
Objection 3: Stopping the usage of slurs is not going to change the culture.

We need to understand that causes in real life are extremely complicated. Most effects that we experience are not caused by "necessary" conditions and neither by sufficient conditions. Driving rashly is not going to necessarily cause an accident and driving safely is not a sufficient condition for avoiding accidents. Real world effects have a lot of causes and associated complexity.

Hence we need to understand claims just like the speaker example. The evaluation should be done based on probabilities, similarity of circumstances, and how presence and absence of the cause in right context makes a difference.
[+] 4 users Like Kanad Kanhere's post
Reply
#2
This should probably be on the blog, methinks.
Reply
#3
(23-Apr-2013, 12:01 PM)Kanad Kanhere Wrote: 1. Caste based discrimination leads to restricted growth opportunities
Objection 1: But Dr Ambedkar accomplished a lot even when there was a lot of caste based discrimination.

2. Misogyny in Indian culture leads to rapes
Objection 2: But this happens even in west.

3. Slurs lead to bigoted culture.
Objection 3: Stopping the usage of slurs is not going to change the culture.

I think some of the objections can be avoided by modifying the claims from "X lead(s) to Y" to a less emphatic construction "X can lead to Y".
Reply
#4
One interesting coinage that helps avoid the traps like the fallacy of the single cause and the fallacy of necessity, is nexus causality. The following article is a useful aid to promote a better understanding of causality, notwithstanding its daunting title.
Nexus causality, moral warfare and misattribution arbitrage

The notion of nexus causality is useful in several topics discussed here:
eg.
- While there is a tendency among some liberal commentators to ascribe acts of violence by fundamentalists wholly to deprivation and othering, religious motivations are very much part of the causal nexus and far from simply coincidental. (Related discussion)
- Suggesting that an average man and and average woman under the BPL category are equally disadvantaged ignores the causal nexus of gender and caste that results in intersectional marginalization. (Related discussion)
[+] 3 users Like arvindiyer's post
Reply
#5
Another objection, that I often come across, is around localization of causality to individuals, even when the causal ownership is attributed to a system.

For e.g. I had a discussion with one of my friends on causal link between violent video games and increase in random teenage violence. He cited articles that concluded with skepticism on the causal link. The problem with those articles was that they used a very restrictive meaning of causality i.e. they used the following parameters to check for a causal link
1. Did a teenager, who frequently played violent video games, commit violent acts more often than others.
2. Did a teenager, who committed some random violent acts, played violent video games.

This is trying to deduce a direct causal link, rather than the indirect systematic causation which is typically claimed. The claim is that exposing teenagers to violent video games desensitizes them towards violence, which leads to a culture where violent acts are on the rise. This subtle distinction is missed more often than not even in the case of culture of misogyny, culture of homophobia etc.
[+] 1 user Likes Kanad Kanhere's post
Reply
#6
(08-May-2013, 06:00 AM)Kanad Kanhere Wrote: 1. Did a teenager, who frequently played violent video games, commit violent acts more often than others.
2. Did a teenager, who committed some random violent acts, played violent video games.

This is trying to deduce a direct causal link, rather than the indirect systematic causation which is typically claimed.

Not sure those two examples are trying to deduce a causal link. But don't such studies atleast have the capability to establish correlation? The causal link might just be a hypothesis which has to be tested using more clever studies after correlation is established.
Reply
#7
(08-May-2013, 06:48 AM)Captain Mandrake Wrote: Not sure those two examples are trying to deduce a causal link. But don't such studies atleast have the capability to establish correlation? The causal link might just be a hypothesis which has to be tested using more clever studies after correlation is established.

I agree that those studies are capable to establish positive correlation, but only at individual levels. What matters is correlation at system level i.e. something like "countries with more teenagers playing violent video games vs countries with less teenagers playing violent video games". That can be complicated because of so many other cultural aspects.

The point was to show how those studies don't really debunk the claim. The claimed causation is Violent-Video-games -> Desensitization -> Culture-that-trivializes-violence -> More-violent-acts. This can't be reduced to individual levels and hence the studies checking the 2 points that I have mentioned earlier don't help.
Reply
#8
(08-May-2013, 08:37 PM)Kanad Kanhere Wrote:
(08-May-2013, 06:48 AM)Captain Mandrake Wrote: Not sure those two examples are trying to deduce a causal link. But don't such studies atleast have the capability to establish correlation? The causal link might just be a hypothesis which has to be tested using more clever studies after correlation is established.

I agree that those studies are capable to establish positive correlation, but only at individual levels. What matters is correlation at system level i.e. something like "countries with more teenagers playing violent video games vs countries with less teenagers playing violent video games". That can be complicated because of so many other cultural aspects.

Yes the differences in cultural aspects will make it difficult to make an apple to apple comparison. But what do we hope to achieve by that comparison? After all American teenagers are going to play their games and commit (or not) their violent acts in America and Australian teenagers are going to play their games and commit (or not) their violent acts in Australia. Just trying to find a relationship between these phenomena with in a country/society might be possible and can be useful.

(08-May-2013, 08:37 PM)Kanad Kanhere Wrote: The point was to show how those studies don't really debunk the claim. The claimed causation is Violent-Video-games -> Desensitization -> Culture-that-trivializes-violence -> More-violent-acts. This can't be reduced to individual levels and hence the studies checking the 2 points that I have mentioned earlier don't help.

Yes. Those studies do not have the capability to test the claim about causation or correlation even if we are only talking about an individual. But such studies can be improved by seeking answers to these questions about a population living in a sufficiently similar culture/society.

Quote:What % of teenagers in the US who commit violets acts have a history of playing violent video games? What % of teenagers in the US play violent video games? Is there a statistically significant difference in these %s?

Data on video game consumption (eg. time spent online playing games) and crime statistics should be available. So answers to these questions must also be available. The answers can be valuable in that they can at least confirm whether or not a correlation exists. If there is no correlation then we can conclude that causation do not exist.

But if we find that correlation exists then we will have to devise clever experiments or painstakingly mine available data to test if there is a causal link. And if a causal link is established we will then have to further devise more intricate experiments to test out the hypothesized causal pathway (Violent-Video-games -> Desensitization -> Culture-that-trivializes-violence -> More-violent-acts).

I understand that your posts are really not about violent video games and violent acts. As per my interpretation of your posts the points (correct me if I am wrong) you were making are the following.

1) In trying to find causal links between two phenomena in social contexts we should not be relying on data about individuals.
2) We should instead be looking for system level causal link but then when we take a system level approach there are so many confounding factors that it will be difficult to test if there is a causal link.

I agree with you on 1). And on 2) though I agree that the system level problem is made difficult because of confounding factors it may not be unsolvable. We may be be able to break the problem into smaller parts and get an answer to your question about causation subject to some simplifying assumptions and caveats.
[+] 1 user Likes Captain Mandrake's post
Reply
#9
(09-May-2013, 08:43 AM)Captain Mandrake Wrote: 1) In trying to find causal links between two phenomena in social contexts we should not be relying on data about individuals.
2) We should instead be looking for system level causal link but then when we take a system level approach there are so many confounding factors that it will be difficult to test if there is a causal link.

I agree with you on 1). And on 2) though I agree that the system level problem is made difficult because of confounding factors it may not be unsolvable. We may be be able to break the problem into smaller parts and get an answer to your question about causation subject to some simplifying assumptions and caveats.

I agree on all counts. I would also like to add is that for 2) step-wise data could be sufficient to imply cultural level causation. For e.g. if there is data to show violent video games lead to desensitization, and then there is different data to show that cultures with higher levels of indifference towards violence lead to more acts of violence, then we have sufficient evidence to claim "violent video games lead to increased violent acts".
Reply
#10
(24-Apr-2013, 01:34 AM)arvindiyer Wrote: One interesting coinage that helps avoid the traps like the fallacy of the single cause and the fallacy of necessity, is nexus causality. The following article is a useful aid to promote a better understanding of causality, notwithstanding its daunting title.
Nexus causality, moral warfare and misattribution arbitrage

This recent BBC news item illustrates one important contemporary instance of nexus causality that is relevant worldwide, while examining the flipside of using slogans like "Every 10 seconds, a child dies of hunger."
Does a child die of hunger every 10 seconds?

Quote:The fact that poor nutrition is identified as an underlying cause of death means that there's also some double counting going on. When you hear that one child dies every few seconds from water-related diseases, for example- or from poverty - some of these children will be the same ones that are said to be dying every few seconds from hunger.
Reply
#11
I have also observed that lack of naturalistic perspective leads to a lot of resistance to the acceptance of influence of culture. The illusion of free-will is compelling. Additionally it takes a certain amount of rebellion in an individual's part to become a freethinker in our present society. This leads to overemphasis on individual choices and underemphasis on cultural causation.

Hence, in my experience, in lot of discussions, there is this suppressed premise that "if individuals change, the culture will change. Hence its individual's fault". This view is not logically unsound, but it is disconnected from reality. We are caused beings and culture has immense role in making us who we are. This reductionism of society to collection of individuals is not helpful. Additionally there is great benefit for individuals to maintain status quo [please refer system justification theory]. All this means that dismantling unjust cultural systems is of greater value than demanding people to "just change".

To explain the above I would like to use example of Sati system.
Sati was prevalent about 300 years back. And now it is banned. And not just that, but considered totally evil by most. Then how could such an evil system be practiced back in those days? There is no reason for us to believe that biology of Indians has changed over a period of couple of centuries, resulting in us being *more ethical*. There is no reason to believe that conscience of Indians got drastically better in a span of decades. What is more likely is that the works of Raja Rammohan Roy and his peers, against the practice of Sati, lead to a cultural change and eventual ban of the system.
Reply
#12
(26-Jun-2013, 04:16 PM)Kanad Kanhere Wrote: I have also observed that lack of naturalistic perspective leads to a lot of resistance to the acceptance of influence of culture. The illusion of free-will is compelling. Additionally it takes a certain amount of rebellion in an individual's part to become a freethinker in our present society. This leads to overemphasis on individual choices and underemphasis on cultural causation.

One way of illustrating the interplay between both top-down (policy-initiated) and bottom-up (aggregated-collective-preference-based) influences is to consider the history of reorganization of Indian states on a linguistic basis.
(i) A spoken language evolves or 'speciates' by a process of 'mutual accommodation of speakers' in local communities, a process that is largely 'emergent' and bottom-up. Influential literary figures (really, members of literary movements) can establish status as a literary language for contemporary forms of a spoken language. Cases in point are the role of patronage structures as well as the agenda of progressive groupings (that maybe treated as more top-down influences) in the development of modern Kannada literature and Telugu literature. In other words, it was not just a coincidence of changing individual reading tastes that shaped the use of written language in late colonial India.
(ii) In independent India, we might say that state boundaries made a Kannada-speaker more Kannada or a Marathi-speaker more Marathi, besides the bottom-up influence of a given language being their mother-tongue. Pending state boundaries, linguistic communities merge into each other and floating populations are casually multilingual to this day. Administrative boundaries enforce a sort of linguistic consolidation which being raised speaking a mother-tongue alone does not seem sufficient to accomplish, through measures like establishment of schools with the defining language as the medium of instruction. It would be reasonable to assume, for instance, that there is more Kannada spoken in Belgaum today thanks to its appropriation into Karnataka and resultant administrative decisions, than would have been if there were no linguistic states.

Here are some earlier discussions about how inadequately acknowledging either top-down or bottom-up influences can adversely affect understanding of social phenomena and the quality of proposed solutions:
(i) State response to riotous behavior will be inadequate if unmindful of social mobilization and treated simply as a coincidence of individual delinquency and hooliganism.
(ii) Undue confidence in legislative silver bullets (eg. Jan Lok Pal) at the cost of slackening attitudinal and behavioral change on part of individual citizens, can result in such initiatives lapsing into an unrealistic transcendental institutionalism.
(iii) The same person can be harmed as an individual (say by a pickpocketing incident at a fair) or harmed as part of a society (due to unreasonable tax policies of the administration). Therefore there is relevance for two ways of doing good, through one-off acts of individual assistance as well as through advocacy and activism for social change, and to eschew one in the name of the other will leave everyone with a feeling of 'something missing'.
[+] 3 users Like arvindiyer's post
Reply


Possibly Related Threads...
Thread Author Replies Views Last Post
  Understanding people's behavior in different cultures Kanad Kanhere 1 5,532 02-Mar-2014, 07:04 AM
Last Post: arvindiyer



Users browsing this thread: 1 Guest(s)