The information is extracted from the Internet; Pan Suiming | Department of Sociology, Renmin University of Chinese
How has life been tampered with as data?
-- Big data is applied to study the "original sin" of human beings
✪
(This article was originally published in New Horizons, Issue 3, 2016, the author Pan Suiming was the director of the Institute of Sexual Sociology and a professor of the Department of Sociology of Chinese People's University, mainly engaged in sociological research methods, sexual sociology and gender anthropology research, known as "the first person in Chinese sexology")
In recent years, there has been a blind worship of big data, that is, following the trend and infinitely elevating without questioning and reflection. This article does not deal with big data and its applications in any natural sciences, but only discusses one fundamental question: can big data be applied to the study of human beings?
Although there have been some doubts about this issue in mainland academic circles, not only are they scarce, but they are also mainly debated at the level of operability in terms of theory, and they have not hit the nail on the head. In fact, the most questionable thing about big data is neither its definition, nor its function or significance, nor the methodological level of "replacing causal analysis with correlation analysis", but the core slogan and basic theory of "everything is quantifiable". It expresses the three meanings of big data: first, without quantification, there is no data, let alone any big data; second, the material world can of course be quantified, but if it is limited to this, then the so-called big data is only an increase in quantity, the nature has not changed, pure hype, for example, weather forecasts have been analyzing massive data, but they do not claim to be big data, let alone form a worship; third, the reason why big data is worshiped now is actually only one point: human behavior and its results are also quantified, and it is claimed to be all-encompassing。
In this way, the nature of big data has changed, from the intrusion of natural science to the study of humanities and social sciences, and from science to "scientism". This is not just a question of research tools, but a fundamental question of epistemology. This article will not list them all, but only analyze what happened to the infinitely rich life practice of human beings in the process of being transformed into "data" by "scientism" at the operational level.
▍ Real life is cropped
Big data admirers tout the "4 Vs" (scale, variety, speed, and value), but they deliberately avoid a fundamental question: In the beginning, what are you collecting is data that can be analyzed?
In sociological surveys, it is possible, but in so-called big data, it is absolutely impossible. Because big data is not the human behavior and its results that researchers actively collect, but a variety of so-called "objective records", which are the "available information" of a very small part of human life, such as traces left by Internet activities, surveillance records, etc.
However, it is well known that of all the information available about human activity, far less is available than it is not. The latter is most typical of all the information of human mental activity, which in the foreseeable future will remain not only unavailable, but also simply unmonitorable. In this way, the information obtained by the so-called big data is, firstly, extremely one-sided, secondly, rambling, thirdly, fragmented, and fourth, meaningless, and cannot be directly used for any quantitative analysis.
How, then, can such information be transformed into analyzable data? First, it must be "defined", i.e., what is kept and what is discarded, second, it must be "classified", i.e., what belongs to what, third, it must be "defined", that is, it is given a specific human meaning to a certain type of information, and finally, it must be "assigned", that is, different definitions must be converted into calculable values.
For example, how can big data producers define whether the traces are intentional or unintentional, idle or staring, actively sought or guided?
Why is "active" defined as "demand"? Finally, what is the value of "demand"? What is the quantitative relationship between the different assignments from "no demand" to "strong demand"?
Obviously, in the process of this tetralogy, it is entirely the researchers themselves who subjectively, artificially and forcibly "sort" the "available information", and completely tamper with the traces of human life into "data" acceptable to their own world view and values. At best, this is called inevitable processing, at worst, it is blatant falsification.
That is to say, the so-called big data does not go beyond the original limitations of "small data" at all: cutting life and tearing life apart, and the "people" who survive as a whole are regarded as a mess of fragments. In real life, human beings are definitely not, and cannot be, "quantitatively" perceive and act in this way. Therefore, big data is not actually to help human thinking, but to replace and control human life experience, which is a bad omen for artificial intelligence.
▍ Social context is ignored
Some people have discovered that big data records the behavior of a single individual, and cannot discover the relationship between different actors; so the question arises: In this real world, is there really a kind of personal behavior that has nothing to do with others? Isn't all the behavior of an individual in a certain interpersonal relationship that can only produce and bring about a certain result?
Society is not a simple collection of individuals, but people are organically organized through various relationships. At the same time, people make various behaviors in a specific social environment, and it is impossible to be imaginative and alone. Therefore, all traces left by human activities must contain infinitely rich social content. If it is discarded, then no matter what kind of data it is, it is not only a glimpse, but also a blind man touching an elephant.
In particular, each person grows into "this person in the here and now" little by little in a particular society, and only then does he or she engage in this behavior. This is the process of constructing the social history of each person, the most important of which is all our social background, living conditions and upbringing.
However, all this often exists only in our own experiences and memories, and it is often difficult to express and even more difficult to record. From the perspective of "objective monitoring", it is simply "the wind passes without a trace". So, even if there is no privacy, even if monitoring can be found in the world, where does the information source of the so-called big data come from? Therefore, for understanding human life, big data is actually a castle in the air.
As mentioned above, this criticism is still very pertinent: "Data doesn't understand social and context, and it creates more noise and misses out what is really valuable." Big data can't solve big problems. ”
▍The main structure is obliterated
"Big data admirers" probably don't know, or dare not admit it, that there is another phenomenon in human life called "subject construction". That is, people's interpretations of their actions are likely to be very different from or even contrary to those of the monitors. The most common thing is that all interpersonal misunderstandings stem from this.
First of all, taking the big data of online shopping as an example, even if you collect all the traces of the Internet and see them digitally at a glance, how do you know that this is what people really think? This kind of "objective measurement" is very close to mineralogy, but people have subjective will, how do you monitor it? Even physics has a "uncertainty principle", not to mention people's subjective will? And secondly, did you know that humans still "present themselves"? To put it mildly, it's performance. If the law does not even accept the results of the polygraph, then how can you screen out the performances of the subjects being monitored? Third, don't you ask the subjects to be monitored to verify them? In judicial trials, even criminals with conclusive evidence must listen to what they have to say in order to make a correct judgment. However, big data admirers simply ignore the existence of the subject's will. Fourth, have you heard of Freud? Do you know that in addition to "motives," there is also "unconscious"?
Even if some data obtained by asking the other party about their wishes still have the question: does the other party have enough ability to express their own wishes? We should not forget Freud, we should not ignore the widespread existence of unconscious behavior, and we should not deny that all traces of human behavior, no matter how massive, cannot actually contain and express the meaning of human life. So, if the actor himself doesn't know what is going on, how do you verify it, and on what basis does it determine the authenticity and degree?
All in all, all attempts to understand human beings and their societies through natural science or digital technology are not necessarily failures, but they cannot deny the importance of human "subjectivity", and the result is inevitably to reduce the real life to the fullest.
Therefore, all the "discoveries" obtained by big data are really just some people describing other people's lives. Others are unaware that they have been described, and there is no means to correct it. As a result, big data is nothing more than a new cognitive hegemony built by a group of technocrats, and its admirers are only eager to use it.
▍ The meaning of life is canceled
The history of at least 2,000 years of humanities and social studies tells us that all human behaviors not only contain the meaning of their lives, but also act in pursuit of their meanings. This is the fundamental difference between people and things.
However, all the so-called "recordable traces" of big data cannot contain information about the meaning of the act if the other party's complaint is not obtained. For example, in all online activities, the actor will not express what he is looking for, nor will he be able to express what kind of value and significance it brings.
In the case of data recorded by shopping websites, it is true that it can contain the almost endless traces that tens of millions of people have left unknowingly when shopping, but does this reflect the shopping preferences of these people? Do these people never buy in physical stores again? Do they always show a preference for one thing at a time? Can they get self-satisfaction with every online purchase?
So, how can you be sure that they will make exactly the same choice in different channels, in different situations? If you can't prove it, then your "big data" is just "big garbage", and it's not just wrong.
Traffic surveillance videos, medical records, communication records, etc., are enough to claim to be "big data". However, all these data are only a record of the scattered aspects of people's lives. Therefore, no matter how big such "big data" is, it cannot solve the following series of common-sense problems: First, is there no close relationship between a person's performance in a certain aspect of life and his/her entire personality and life?
Peasant workers are too expensive to eat a 20 yuan box lunch; is it just a so-called "consumption choice" that a rich man buys a car of tens of millions without blinking an eye? Second, is it not the case that all aspects of human life influence each other? Peasant workers can get a gift of 200 yuan when they eat a 20 yuan box lunch; and a rich man who buys tens of millions of cars but refuses to do a little charity, is this just a so-called "buying habit"? Third, is not a person's life constrained by social, cultural, historical, and other factors? The reason why migrant workers have to eat a 20 yuan box lunch is not only because of low wages, but also because their dependency coefficient, unemployment possibility, occupational risks, etc. are much higher than those of the rich. Is this also "disposable funds"?
So, the bigger the data, the bigger the error.
▍Conclusion: Original sin is original sin
Everything discussed in this article is actually derived from the long-standing criticism of "quantitative research" in humanities and social studies. The cult of big data is just the latest manifestation of this trend of thought, but it is just dressed in a more glamorous coat.
In Christian doctrine, original sin is not only innate, but also lifelong, and cannot be eliminated through one's own salvation.
Unfortunately, the same is true for quantitative research. No matter how developed its technical means are, no matter how large its data is, once it is applied to humanities and social studies, its defects and ills cannot be avoided at all, and at best it is only reduced to a certain extent. In the final analysis, the "cult of big data" is actually the doomsday cry after "scientism" has failed in the face of human history. If science has failed to stop Hitler's rule, nor has it been able to predict all the development of mankind since then, then it is by no means a question of "lack of art", but a misuse of it, a crossing of the line and running into a realm that it is powerless.
Of course, this does not mean that quantitative research and big data must not be used, but it expresses three meanings: first, neither of them can question nor replace all kinds of non-quantitative humanities and social studies; second, only quantitative research that deeply reflects on these congenital defects and fully demonstrates them is qualified to retain a place in humanities and social studies; third, the two kinds of research are like two railroad tracks, one is indispensable, but they extend parallel and never intersect.