This article was first published in the new issue of “Reading” in 2021, and is authorized to be reprinted in Tiger Sniff. For more articles, you can subscribe to buy “Reading” magazine or pay attention to the WeChat public number: Reading Magazine (ID: dushu_magazine), by Xu Yingjin, header photo from: “Who am I: No Absolutely Safe System” stills
When the Internet economy first emerged in the 1990s, the media had a certain rose-colored fantasy about the development of the Internet: the Internet would turn the whole world into a whole flatland of information. In this way, the various differences between knowledge-intensive and knowledge-sparse geographies became irrelevant, and quality information was able to flow freely to any corner of the world.
However, over time, the real Internet world has brought about the so-called “information cocoons” and “echo chamber effect” – the Internet has fed in a large amount of information, which has led to a large number of Internet users becoming “silkworms” trapped in cocoons, seeing only their own acres of land in front of them; or becoming lonely speakers trapped in a secret room, hearing only their own echoes.
As a result, different people with different preferences for access to information, but in the current Internet environment has become a dead information island, intensifying the already existing information barriers in the world.
Why is this bizarre phenomenon happening? I believe that Internet technology, and the Big Data technology behind it, takes full advantage of certain features of human cognitive architecture – in particular, the “heuristics” that speed up the information processing process. “The bundling effect between information with the same content attributes and a specific group of people becomes more obvious. The logic behind this bundling mechanism is the inner logic of capital operation.
There is no doubt that no matter how cutting-edge the data technology supporting the current Internet operation is, the majority of Internet users and the operators and controllers of the relevant platforms are still human beings. In other words, based on the biological nature of human beings, the current operation of the Internet is actually distinctly “semi-biological”.
However, it is intriguing that the mainstream Internet narrative tends to turn a blind eye to this obvious fact, and is more accustomed to simplifying a half-biological history of Internet evolution into a purely “silicon-based” history of technological evolution. Our task now is to recover this oversimplified picture.
From the point of view of evolutionary psychology, the human biological brain, like the brains of other animals, can be considered a sort of abstract biological computer – designed to extract useful information from the complex external environment in order to improve the reproductive fitness of the organism itself. However, the real operating procedure of this “computer” highly reflects the principle of “frugality”, that is, the use of minimal energy loss in information processing to produce those behaviors that can best improve reproductive fitness.
The following are some well-known cognitive effects that fully embody these “frugality” principles – each of which brings survival benefits to the organism, but also cognitive biases.
One is the anchoring effect. This effect says that the first impressions acquired by a cognitive subject acquire a more solid position in its memory like an anchor, and make it difficult for later impressions of the same type of thing to play the same cognitive role as the original impression. This effect can obviously save the cognitive subject’s information processing cost.
For example, if a subject wants to determine what color a crow is, he may not have to spend his energy examining all the crows in the world, but only the first ones he sees. However, because of this effect, the cognitive subject is prone to make the mistake of “hasty generalization”.
Second, the labeling effect. That is, the cognitive subject will label the various things he sees in order to facilitate his classification of everything. The advantage of this is that it obviously reduces the number of information modules to be processed by the subject and reduces the subject’s management cost, but the disadvantage is also obvious: if the subject classifies things according to oversimplified (or wrong) labels, the subject will have a wrong cognitive picture of the world, and will thus hit a wall in the real world.
Third, the “easy thinker wins” effect. This effect means that when a subject judges whether a certain attribute of something is obvious or not (e.g., whether the divorce rate of celebrities in the entertainment industry is high or not), it is mainly based on whether the attribute is easily evoked in the mind rather than on how often it appears in the real world. Considering the large cognitive load required to retrieve the “true frequency of occurrence of the attribute”, the existence of this effect obviously reduces the subject’s cognitive load significantly. However, in some cases, the subject may be led by the nature of things that come to his mind by chance, and thus lose his grasp of the true ratio between different properties of things because of this effect.
Fourth, the extreme expectation effect. That is, when the subject is faced with a state of affairs that is quite related to its interests is not very clear, the subject will hastily make a polarized expectation of its final direction with “good” and “bad” colors. The advantage of this approach is that it allows the subject to prepare a response to the end of the matter before it finally appears, shortening the chain between “information processing” and “action” and thus reducing the cost of information processing. The disadvantage is that it makes the subject lose the ability to respond to the drastic change of the situation because of the rash bet.
Fifth, the float effect. The effect is that when the subject is unable to judge the truthfulness of the information according to his own cognitive ability, he tends to draw a conclusion according to the way the mainstream society judges the issue, just like jumping on a seemingly fancy float when participating in a parade. The advantage of this is that the cognitive subject can reduce the cost of information processing by “borrowing” from the social mainstream or relevant authorities; the disadvantage is that the cognitive subject is likely to be misled by some false authorities and fall into the cognitive trap.
It should be noted that although each of the above effects brings both advantages and disadvantages to the cognitive subject’s adaptability, they are outweighed by the associated disadvantages during the long gathering-hunting era that allowed the evolution of human cognitive architecture.
Because the primitive gathering-hunting environment is naturally occurring, there are no obvious signs of human design. It does not lead to serious problems of misattribution. Moreover, the environmental elements of the primitive gathering-hunting era were relatively simple, and the cognitive structure of primitive people only needed to focus on those environmental factors that were highly relevant to their biological needs (such as food, mates, natural enemies, natural disasters, etc.), without considering those abstract concepts (such as money, capital, etc.). This makes the “labeling effect” less likely to lead to errors.
At the same time, in such environments, the types of events tied to the “extreme expectation effect” are already highly restricted to basic material needs. In contrast, social polarization due to abstract ideological struggles was not common at that time. At that time, mediated information dissemination was also very underdeveloped, and the frequency with which a person could think of something carrying a certain nature was often the objective frequency with which he saw that thing carrying that nature, rather than the frequency with which he saw that thing carrying that nature under the deliberate guidance of some third-party medium. This situation makes the negative impact of the “easy thinker, easy winner effect” not too great.
Similarly, because of the underdevelopment of socially mediated media, any cognitive subject can easily find the source of information by tracing the chain of information dissemination and hold the sender of the information as the source socially responsible for the quality of the information itself. This also curbs the negative implication of “float effect” to a considerable extent.
In the modern information society, however, the environmental parameters of the gathering-hunting era have been comprehensively modified. This has made the mental characteristics that were adaptive in the original environment less adaptive.
First, the environment we live in has long been highly artificial. For example, the advertisements you see in the subway are not naturally formed, but artificially designed; the girls you see on the road are wearing a certain color not because it enhances their biological adaptability in a particular environment, but simply because it is the “trendy color” of the year (for The mechanism of setting the “fashion color” is itself part of the modern consumerist discourse).
In the virtual world of the Internet, this “artificiality” has been exaggerated to the point of no return. Compared to the construction of physical artifacts, the cost of constructing digital virtual objects is lower, and therefore, the less rigid constraints from the physical world are imposed on the setting of these virtual objects. This makes the “anchoring effect” more likely to lead to cognitive bias, because the initial impressions that subjects receive about something on the Internet may be the result of artificial arrangements.
Second, modern humans are in a complex hardware and software environment, and to cope with this complexity, humans have to develop very complex abstract concepts to operate the relevant social machines. But since the basic elements of human’s actual cognitive ability (such as brain capacity, working memory capacity, etc.) are not equally complicated, this makes the gap between the conceptual map that a single modern human can actually master and the complex environment sharply widened. This makes the negative connotation of “labeling effect” more likely to be demonstrated in modern society.
Again, the information intermediation in modern society is so developed that a person’s knowledge of news about a distant country is limited to the information interface presented to him by the media. In the premise that the information source itself is polluted, the above situation will open the door for the full proliferation of the “float effect”. Likewise, the negative implication of the “easy thinker easy winner effect” will be expanded, because the attributes of things that the information subject is more likely to recall are not the real frequency of the relevant attributes of things, but the frequency of the relevant attributes in the media.
Finally, the material life of modern society is highly developed, and a large number of subjects can have more leisure time to engage in mental activities, therefore, the range of events to which the “extreme expectation effect” in the primitive mind applies is also sharply expanded, i.e., from the “first world” (physical world) and “second world” (mental world) as Popper said, to the “third world” (mental virtual world) rapidly.
In other words, the number of causative events that lead to polarization is rapidly expanding, and the possibility of tearing people apart has increased significantly. In the digital era, as the cost of creating false information to disrupt the information field of the “third world” becomes smaller and smaller, more cognitive subjects are more likely to be affected by the aforementioned “float effect” and “anchoring effect In the digital era, the cost of creating false information to disrupt the information field of the “third world” has become smaller, and more cognitive subjects are more likely to be compelled by false positions under the effect of the aforementioned “float effect” and “anchoring effect”, thus widening the gap with others.
Color authority PANTONE (Colorcom) publishes the color of the year every year (source: Adobe Stock)
All these above-mentioned changes in environmental elements give rise to information cocoons in the Internet era. The causal mechanism behind them can be broken down into the following steps.
Step 1: Different virtual information producers will generate different value judgment propositions according to their different interests, such as a positive judgment on a certain positive attribute of a commodity. Therefore, the virtual world becomes a battlefield where these massive value judgment propositions (and the accompanying artificial body of evidence) compete with each other.
Step 2: For a single information recipient, the “thrifty” nature of its cognitive structure drives it to randomly select a small number of judgments from the vast amount of information as part of its belief system to avoid “information overload” in its cognitive structure. The problem of “information overload”. And in the online world, an important indicator of someone’s adoption of a belief is to express itself in terms of browsing records, i.e., he clicks and views the virtual content interface (article, audio or video, etc.) that expresses that belief.
Step 3: The content recommendation technology driven by big data technology will recommend similar information to individuals based on their information trajectory in the virtual world, so as to use the anchoring effect to quickly form the information retrieval habits of relevant subjects to facilitate the future development of the advertising push business. Step 4: In this way, the differences between the starting belief systems formed by individuals due to certain contingent factors will be continuously expanded through the collusion between the big data environment and individual cognitive architecture, thus forming an information cocoon.
Faced with such a situation, how to break the cocoon?
Before discussing the way to break the cocoon, we must first determine what matters we cannot change. First of all, it is necessary to affirm that the frugality of human cognitive architecture is a biological constant, and we cannot improve the cognitive situation by systematically modifying the brain, whether for bioethical or technical feasibility reasons. What needs to be changed, instead, is clearly only those data technologies that make it possible to amplify the full extent of the flaws in human cognitive architecture.
Specifically, it seems that we can change the deeper logic of the currently emerging content recommendation technology from the track of “reinforcing users’ existing purchasing habits” to the new track of “expanding users’ interests and increasing their horizons”.
But is this shift possible? This question itself can be broken down into two sub-questions: First, is the technological revolution that will allow this shift to happen possible? Second, can the social resources (such as capital, time and human resources) be made available to enable the above technological revolution to take place?
The author is cautiously optimistic about the first question. Needless to say, with the support of Bayesian statistics and deep learning technology, the mainstream web content recommendation technology nowadays tends to predict the relevant web behaviors of specific users based on the existing web trajectory characteristics of most people, and this approach also tends to make individuals become impersonal footnotes of the established “user profile”. Nevertheless, reverse technology development is not impossible.
For example, it is certainly possible for a technology developer to push content products to users that are contrary to their browsing habits, such as videos for Democratic voters in the United States that reflect the Republican Party’s political philosophy. But why do so few people do this in the real world? This brings up the question of the capital environment that supports the development and use of content recommendation technologies. The development and use of digital technology of course requires significant capital, and capital investment requires a return.
Take YouTube for example: on this platform, the return on capital is often related to the rate of ad viewing of the inserted content video, which in turn is related to the number of clicks on the video itself. Obviously, the only way to maximize the number of clicks on a video with any particular political leaning is to recommend it to other users who might hold that leaning, rather than to users who resist it. In turn, with a relatively established user base, content producers will solidify the content leanings of their products according to their tastes to ensure their own financial gain.
In other words, if someone were to develop a new content recommendation technology aimed at expanding users’ horizons and weakening the information cocoon effect, the technology would be stifled and combated, or at least ignored, by the power of capital at the level of the motivation for its development.
From the above analysis, there has been a high degree of coupling between the operating logic of capital and the way the primitive human mind works, and the path to breaking the cocoon seems to have been closed. But is there any other way to improve the situation?
There is a way. The basic idea is that those in control of the top resources understand this: to increase the attention of content, it is not necessary to strengthen the barriers of the cocoon – sometimes it is beneficial to capital to weaken this barrier instead.
In medieval Europe, where groups of people dominated each other, no count or duke could directly mobilize the human resources in the territories of other counts or dukes. In this case, the whole Europe was a highly gridded “resource cocoon”, that is to say, each feudal lord could only access the resources of his territory, but not the resources of others.
Nonetheless, annexations between lords continued to occur, and voluntary consolidation of resources between lords was a frequent occurrence throughout history. Such consolidation relied on three main approaches.
Direct marriages between vassals, alliances under the bonding of religious power (especially that of the Holy See), and alliances under the bonding of symbolic high power (such as that of the Holy Roman Emperor of Germany) in the form of the “Son of Zhou”. However, regardless of the method of resource integration, the purpose is to solve the problem of resource dispersion caused by the isolation of each “resource cocoon”, in order to revitalize certain idle resources and ultimately achieve the maximum benefit of resource utilization.
A map of medieval cities in Europe (source: pin.it)
In a sense, the “information cocoon” in today’s Internet world can be seen as a virtualized version of the medieval “resource cocoon”; and certain groups of fans organized according to certain interests can be compared to the human resources of a medieval lord’s domain. all the human resources available in the territory of a medieval lord.
Therefore, just as it was possible for the medieval lords to weaken the barriers between the “resource cocoons” under the dictates of certain common interests, it is also possible for the capital controllers of today’s Internet era to weaken the barriers between the “information cocoons” under the consideration of certain common interests in order to achieve the goal of “information cocoons. in order to achieve some long-term benefits.
A particular illustration of this possibility is the phenomenon of “out-of-circle”, which is now on the rise. In short, an Internet label personality with significant IP value who has accumulated a large number of followers in one circle can also achieve significant success in another circle, thus allowing the new fan base to merge with the old one. This fusion happens precisely because the “out-of-circle” phenomenon itself magically exploits the audience psychological mechanism that originally caused the “information cocoon” effect, and achieves the effect of breaking the cocoon barrier. The specific operation process is as follows.
The popularity accumulated by the idol itself makes the idol’s behavior itself a giant float, attracting a lot of online attention; the “cross-circle” or “out-of-circle” behavior of the idol makes the idol’s labels diversify, thus objectively enriching Once the cross-circle behavior of idols becomes a stereotype, the anchoring effect of the audience will be activated, thus creating expectations for the cross-circle activities of other idols and forming an invisible pressure to contribute to the elimination of the best and the worst in the idol world. Once the cross-circle behavior becomes a certain momentum, the operation of the polarization effect will also shift – what was originally the polarization and contempt of the respective fan circles between one circle and another circle now becomes the contempt of the fan base of idols who can cross the circle against the fan base of idols who cannot cross the circle. It is obvious that this kind of disdain is caused by the IPs themselves.
Obviously, this change in audience psychology caused by the IP’s own autonomous drift does not happen in a forced way, and is therefore more wonderful than forcing users to recommend information “outside the circle” with software. However, in order to achieve this wonderful effect of destroying the cocoon itself with the psychological mechanism that creates the cocoon itself, two conditions must be met.
First, the top resource controllers, need to be strategically aware of the long-term economic benefits that can be generated by breaking the cocoon. From the perspective of permutation, if the three IP resources at the disposal of a company originally had only A, B, and C traits each of the traffic, then the way these traits are arranged and combined with each other will generate a large number of new possible ways of presenting IP, thus leading to a comprehensive enrichment of content resources.
Second, icons need to improve their own quality in order to provide relevant skills and knowledge reserves for loading more IP attributes. In the above process, the government can promote the joint cultural creation program between the disadvantaged cultural labels (such as traditional opera and traditional crafts) and the strong cultural labels (such as secondary yuan culture) in a “silent” way, so that some disadvantaged cultural labels can ride on the “big float The “Big Flower Car” is a creative program that allows disadvantaged cultural labels to ride on the “Big Flower Car”.
Posted by:CoinYuppie，Reprinted with attribution to:https://coinyuppie.com/human-cognition-and-the-information-cocoon/
Coinyuppie is an open information publishing platform, all information provided is not related to the views and positions of coinyuppie, and does not constitute any investment and financial advice. Users are expected to carefully screen and prevent risks.