[author]Liu Sida

[content]

Breaking the superstition of numbers: On the "pseudo scientificity" of social science and law

*Author Liu Sida

Professor at the Faculty of Law of the University of Hong Kong and Vice Dean of the Institute of Chinese Law and Society at Shanghai Jiao Tong University

Abstract: Looking back at history, it is found that from the 1990s to the early 21st century, social science law research mainly adopted qualitative research methods such as participant observation and interviews. However, since 2013, there has been a clear "scientific turn" in social science law. With the emergence of judicial documents online and various statistical data, scholars engaged in legal empirical research have become superstitious about various statistical data and numbers. However, many issues that rely solely on data for research are often overlooked. Firstly, the disclosure of statistical data is not complete; Secondly, relying on data cannot provide a comprehensive understanding of legal facts and the context of events; Thirdly, excessive reliance on data can lead to academic research becoming increasingly dependent on state power and resources, resulting in a loss of independence. Therefore, to break the superstition of "numbers", on the one hand, breakthroughs can be sought in research methods, making full use of new research methods and information sources brought by technology. Online videos and online ethnography are good methods; On the other hand, we need to tell good stories, handle the relationship between theory and description, and conduct research that can touch people's hearts.

Introduction

Today, I will talk about a topic that is quite academic, called "Breaking the Digital Superstition - On the 'Pseudo Scientificity' of Social Science and Law". Why am I talking about this topic? It is because social science law has become a research field in the Chinese legal community that, although not necessarily mainstream, at least has its own place. This is actually a relatively recent event. The last time I held the Law and Social Sciences Annual Conference in Kunming was in 2013, and I gave a lecture in the same room. At that time, social science law was still in a relatively weak position in the legal field. I was about the same generation as young and middle-aged scholars such as Hou Meng and Chen Baifeng who conducted research in social science law. At that time, this field was not called "social science law", but "law and social sciences". I often joke with them that with just a dozen or so people and seven or eight guns, we have actually created a 'law and social science'. Later on, domestic peers held multiple dialogues with legal doctrine to establish the theoretical and methodological legitimacy of social science law. Professor Hou Meng and his team did a lot of foundational work. Ten years later, this kind of dialogue is no longer meaningful because social science law has become a "legitimate" field in China, and various mainstream legal journals often have works on legal empirical research. This is a significant improvement compared to the situation ten years ago.

However, in recent years, I have discovered a new problem, which is the "number superstition" that I want to talk about today. Why is it called 'number superstition'? Because social science and law are becoming increasingly like 'pseudoscience'. After graduating from Peking University Law School with a bachelor's degree, I went to the University of Chicago to study sociology, so I often joke that I no longer understand the law very well and belong to the category of half "legal blind". Moreover, I am not a scientist, but a "pseudo scientist". My college classmates and I usually call ourselves "law blind", and my high school classmates usually call ourselves "pseudo scientists". I graduated from the Mathematics Experimental Class of Renmin University Affiliated Middle School. The vast majority of my high school classmates studied science and engineering, and they are the real scientists and engineers. I am a complete "pseudo scientist". At that time, I was half joking, but after more than a decade, I increasingly felt that my research field - whether it was called "legal sociology," "law and social sciences," or "social science law" - was becoming more and more like a "pseudoscience.

Why did this transformation occur? This transformation probably occurred in 2013, during the ten years after the annual conference of law and social sciences in Kunming, there was a significant shift in legal empirical research. In fact, I wrote about this issue in a 2016 article titled 'The Rise and Criticism of the' Law and Social Movement 'in the United States', with the subtitle' On the Future Direction of Chinese Social Science and Law '. The conclusion section of this article proposes several possible trends that I observed at the time in social science law, the first of which is that social science law is likely to be undergoing a "scientific turn". I have already seen this trend in 2016.

1. The 'Scientific Turn' of Social Science Law

If we look at the history of Chinese legal sociology, in the 1990s, Su Li used qualitative research methods to study local resources and sent law to rural areas. He conducted research through interviews, films, and historical documents. Including young scholars Qiang Shigong and Zhao Xiaoli at that time, their research on mediation and rural dispute resolution also used ethnographic methods such as participant observation and interviews. The generation of Chen Baifeng, Dong Leiming, and Di Jinhua, who were cultivated under the tradition of the "Huazhong Village Governance School," were the backbone of social science and law, and also engaged in ethnographic research. For example, Professor Chen Baifeng's doctoral thesis, published as "Rural Jianghu - Research on 'Gangsters' in the Two Lakes Plain" (hereinafter referred to as "Rural Jianghu"), I believe it is still the highest level of research in contemporary Chinese legal sociology and the pinnacle of this field. But if you read this book, you will find that it is purely an ethnographic study without any numbers. It is a research formed by Professor Chen's long-term observation in a village.

So until about 10 years ago, the mainstream of research in social science and law was actually qualitative methods, such as interviews, participant observation, and historical literature analysis. There are certainly many reasons for this, and the most direct reason is that there were not many numbers available at the beginning, and there were no good statistical data from the 1990s or even the early 21st century. Additionally, another reason is that legal scholars were not very knowledgeable about numbers in the past, making it impractical for them to conduct complex statistical analyses. Because many legal scholars basically stop studying mathematics after the age of 20, it is quite difficult for them to engage in the so-called "data law" and "digital law" nowadays. So, traditionally, empirical research on law in China has mainly focused on qualitative research. But in the past decade, there has been a significant shift in social science and law, known as the 'scientific turn'. There are a series of reasons for the emergence of the "scientific turn".

Some scholars in their 40s, who are basically the same age as me, have really started using quantitative methods to conduct empirical research, and they have done a good job. For example, Professor Cheng Jinhua from Shanghai Jiao Tong University studied social sciences at the Hong Kong University of Science and Technology before pursuing his PhD at Yale Law School. He used statistical methods very well and conducted some highly skilled quantitative research after returning to China. There are more and more returnee legal scholars like Mr. Cheng who engage in quantitative research. Some scholars are also locally trained and have started using methods different from qualitative methods to conduct legal empirical research in recent years, such as Professor Li Xueyao, who also teaches at Shanghai Jiao Tong University. His original research interests were similar to mine. His doctoral thesis focused on the study of the legal profession, but later he shifted to the research of legal cognitive science. For example, his study on the impact of cognitive fluency on judicial decisions is an early example of using experimental methods for empirical research in law. Professor Guo Chunzhen from Xiamen University has also conducted similar experiments in legal cognitive science. In addition, some of the earlier generation of legal scholars have also transitioned to empirical research, such as Professor Zuo Weimin and his students at Sichuan University who have conducted a series of empirical studies on courts and criminal justice over the past decade, mainly using statistical methods, and have also been very successful.

I initially welcomed these relatively new quantitative research methods in the Chinese legal community at that time because they were all good attempts. Qualitative methods alone are not the ideal situation in legal sociology, and it is better to have more diverse methods. My viewpoint has always been clear - methods are means, not ends, and we are not 'methods for the sake of methods'. No matter what research is done, adopting a certain method is to explain a specific problem. Some questions may be better observed through participation, some questions may be better interviewed, and some questions may be better surveyed through questionnaires. There is no good or bad method in itself, it just means different methods are used to solve different problems. The utensils used for eating are like the methods and utensils used in research institutes. For example, if you like to use chopsticks, but there is a large piece of steak in front of you, it would be very inconvenient for you to use chopsticks to eat. Of course, you can also eat steak with chopsticks, but when facing a whole steak, eating with chopsticks is definitely not more convenient than eating with a knife and fork. On the contrary, if you have a bowl of Crossing the bridge noodles in front of you, but you have to eat it with a knife and fork, it is not impossible, but it will be very difficult to eat. If you are given a pair of chopsticks to eat rice noodle, you will be able to eat it easily. So chopsticks and knives and forks are not inherently superior or inferior, they each have their own strengths and are means, not ends. The purpose is to make you eat the food in front of you more efficiently and enjoy it more. The same goes for research methods, which ultimately serve the purpose. The method used depends on the specific problem to be studied. Some problems may be better treated with quantitative methods, while others may be better treated with qualitative methods, and cannot be generalized. I have always held this position. So, I initially welcomed the emergence of these quantitative methods, and the diversification of methods is a good thing.

But later on, there were some new changes, the most obvious of which was the increasing number of various "big data". The most frequently used Chinese judicial document website in recent years was launched on July 1, 2013. The online publication of judicial documents is a revolutionary breakthrough, as for many years, scholars like us who conduct legal empirical research have been in a state of 'cooking without rice'. In the past, many people chose qualitative methods such as interviews and observation, of course, there were methodological preferences, but there was also a reason for the lack of quantitative data. Even if your quantitative analysis skills are good, it is still difficult for a clever woman to cook without rice. But this situation has undergone tremendous changes in the past 10 years. Nowadays, big data and artificial intelligence are very advanced, and various numbers are everywhere. Moreover, the cost of statistics - whether it is time cost or labor cost - has decreased significantly compared to before, and many things have been automated.

Give an example that is not in the field of legal research. I have always been a football fan since I was young. When I watched football games as a child, many commentators on TV would say this: Rijkaard passed the ball horizontally to Gullit, Gullit passed it directly to Van Basten, and Van Basten made a powerful shot... They all described the behaviors and events happening on those stadiums. In the past 10 years, I wonder if anyone has noticed that football commentary has undergone significant changes, with various data used extensively, such as "now it's the fifth corner kick of this game" and "this player has received the seventh yellow card of this season"... This kind of thing was impossible before because the previous level of technology could not reach it. But with today's live commentary, these data are right in front of us and can be viewed anytime, making it very easy. The same goes for the legal field, where data has transformed from a state of 'no rice to cook' to a state of 'rice on the streets' in less than 10 years. In the years when judicial documents have been online, the total number of judicial documents that have been online is nine digits, which is a data resource of billions. Therefore, in the past 10 years, legal scholars have rushed forward and conducted a large amount of research on so-called "data law" or "digital law", especially relying on data from China Judgments Online.

2. Problems that arise after the scientific shift

After changing from 'no rice in the pot' to 'rice in the pot', a series of problems arose. Most legal scholars used to be "ignorant of numbers", but now suddenly there are so many data and judgments coming out, and everyone rushes up to "wash rice" and use it to cook, which presents a very obvious problem.

Firstly, rice is not so easy to find. Although there are at least 100 million judicial documents online, this is incomplete because not every judicial document is online. Until now, we still don't know which types of judicial documents are more online and which types are less online. Of course, scholars both domestically and internationally have estimated this, but it can only be roughly estimated. For example, the proportion of judicial documents related to personal privacy, state secrets, and national security in criminal cases that are online is definitely lower than that in ordinary civil cases. But what is the specific percentage of internet usage? And it's not just a matter of percentage, what are the differences in case types between criminal cases where the judgment documents are online and those where the judgment documents are not online? We cannot know about this. So, from a sampling perspective, can using all criminal cases with online judgment documents as a sample truly represent the distribution of types of criminal cases nationwide? It is likely not possible. In some other fields, the representativeness of the sample may be much stronger, such as corporate debt disputes, intellectual property disputes, and other civil and commercial cases, which do not have much privacy and sensitivity, and the proportion of online cases may be higher than that of criminal cases. But even so, it is not 100%. We cannot know what percentage it is, nor do we know if there is a difference in the types of cases between those who are online and those who are not online, and how big the difference is. I believe that every field has differences, but what exactly are the differences? What are the standards for online judgment documents? What kind of case doesn't go online? Until now, we still don't know.

Let's use rice as an analogy. If you have a pile of rice with up to nine digits in front of you, it's not that simple to clean it up. The use of China Judgments Online for big data research has gone through several stages. The first stage is that everyone rushes forward in a swarm, downloading as many cases as possible and then making preliminary descriptions. I often joke that this is just "drawing a big cake", using these numbers to create pie charts that can only reveal some basic statistical distributions, without much academic value. I call this practice 'drawing cakes to satisfy hunger' because I don't know how to analyze, so I drew many 'big cakes'. The first generation of research based on China Judgments Online was almost all about "drawing big cakes", and even now many people are still "drawing big cakes", which is a very low-level quantitative research. After a few years, some scholars who took the data from China Judgments Online seriously realized that "drawing a big cake" was not feasible, because even if all the online documents were treated as a complete database, its representativeness was still problematic. Instead of doing so, it's better to limit the research scope to a specific type of case, such as drunk driving or drug trafficking in criminal cases. This way, the total number of cases and sample size may not be as large, and the number of cases may increase from millions or tens of millions to tens of thousands or even tens of thousands. Compared to millions of samples, analyzing tens of thousands of samples is much easier, and sample bias is likely to be smaller.

In addition, there are also problems with putting national data together, as the proportion of court judgments online varies across the country. For cases of the same type, instead of grouping national data together and drawing a few bar charts or pie charts, it is better to look at Yunnan Province or Kunming City, or even study the judicial judgment of a grassroots court in Kunming City. Perhaps the analysis results will be more reliable. Compared to those who are eager for great achievements and analyze millions or tens of millions of cases, this small sample analysis targeting a certain place or court may be more scientific. This is the first question I'm talking about, from 'no rice in the pot' to 'rice in the pot', 'washing rice' itself has become a problem. This problem has not been solved yet. Some people have solved it better, while others have not even considered the issue of sample representativeness and blindly conducted a lot of statistical analysis.

Secondly, even assuming that rice can be washed relatively cleanly, some scholars at home and abroad now have a relatively high level of rice washing skills, and there are still questions about what kind of rice to cook with this rice. Nowadays, using big data for research, whether it's thousands of judgments, millions, hundreds of thousands, or tens of thousands, the result is basically plain rice without any vegetables. Because the judgment document cannot reflect the full picture of the case. The information reflected in the judicial judgment after following a case from beginning to end is only a very small part of all the information in this case. In recent years, I have interviewed hundreds of lawyers across the country and often heard complaints that they passionately defended themselves in court for half a day, but many of their defense opinions were not included in the judicial judgment. After the release of the China Judgments Online, judicial documents were made public and judges had to write them carefully. Many of the original judicial documents were shorter than they are now, with some even only one or two pages. Many things that happened during the trial of the case were not reflected in the judgment.

One of the most famous sociologists of the 20th century, Erving Goffman, has a classic classification of human activities called "foreground" and "background". Our judicial activities also have a "front desk" and a "back desk". Judicial documents can be said to be the "front desk of the front desk", which solidifies what is publicly available through judicial procedures in the "front desk" into text. The events that occurred during the court hearing that were not included in the judgment documents also belong to the "front desk". Although the court hearing is theoretically public and citizens can enter with their ID cards to listen, it is not so easy to enter the court gate. Even the 'front stage' of the trial is not completely public, so many things that happen in the 'back stage' are even less known. For example, how lawyers, prosecutors, parties communicate with judges, and whether anyone goes to sit in and cause trouble at the entrance of the court... These various things that happen in the judicial process cannot be reflected in the judgment documents. Therefore, when conducting legal empirical research, the focus should not be solely on judicial documents. The most classic legal sociological issues, such as the interaction between lawyers and clients, how judges resolve disputes and mediate, and the formation of legal awareness among clients, cannot be reflected in judicial documents. Why do I say that just looking at big data is like cooking a pot of white rice? It is precisely because of these issues that truly make a study of something flavorful and valuable (a dish). If there were no these "dishes", only white rice, which everyone keeps eating for so many years, the taste of the rice would not have changed much, and it would still be tasteless.

There is a deeper issue, which is that excessive reliance on numbers will make academic research increasingly dependent on the power and resources of the state, which is a worrying thing. In 1998, Professor Su Li published a famous article titled "Power Resources in Legal Sociology Investigation: Reflections on a Sociological Investigation Process". At that time, there was not much quantitative analysis in Chinese legal sociology, and quantitative methods were rarely used. After conducting field research, Professor Su Li had some insights and his observations and feelings were very keen. In this article, he points out that as a researcher, conducting field research, which involves interacting with people, requires the use of a lot of power resources in the Chinese context. In fact, in the context of various countries around the world, there are more or less power and resource issues, but there are still differences. For example, Patricia Ewick and Susan S. Silbey's classic research on citizen legal consciousness. They randomly sampled some communities in New Jersey, USA, knocked on strangers' doors in the communities, and said, 'We're going to do a study on communities now. Can we talk to you about it?'. There are really many people who talk to them, and they have conducted hundreds of interviews in total. But if you do this in most parts of China, it may not be easy to achieve. If you knock on a stranger's door, they may kick you out. This approach is not feasible because Chinese people rely on human relations and relationships, which contain many power resources. For example, if I want to interview a lawyer in a place, who should I go through to find the lawyer? To give a more specific example, I have just started a new research project on the living conditions of young lawyers in China. Because young lawyers are facing many difficulties and challenges nowadays, I want to conduct empirical research to gain a deeper understanding of their living conditions. Assuming my purpose is to interview some young lawyers in Kunming, I have several different ways. I asked a teacher from Yunnan University to introduce me to some of their former students, and I went to interview them. This is one way. Another way is, if I know some senior lawyers in Kunming, I can directly go to a law firm director and ask him to introduce me to the young lawyers in his firm. There is another way, assuming I know the director of the Kunming Judicial Bureau or the president of the Bar Association, I can ask the director or president to arrange for me. But the problem is that using these three different paths, the people I see will be different, and the information I can get will also be different.

When I was conducting field research for my doctoral thesis, I went to 12 different provinces and conducted over 200 interviews. The ways I entered each place were different. Some places were found through personal relationships, while others could not be found through personal relationships, so I had to rely on more formal relationships such as teachers or leaders. The results were different. For example, I once went to a certain county for an interview, and my classmate's relative was the secretary of the county's political and legal committee, so the two lawyers who received me were particularly enthusiastic. They enthusiastically talked to me for two hours, but that doesn't guarantee that everything they said is true. For the most obvious example, those two lawyers told me that in their firm, if the client only comes for consultation, not as a case agent, they never charge the client. I thought it was really good at the time. The interview was held in the conference room on the second floor of the law firm. After the interview, I asked these two lawyers if they could take me on a tour of the firm, and they said it was no problem. However, as soon as we ascended the stairs from the second floor to the first floor, I saw a table at the entrance for receiving clients. There was a sign hanging on the table with four big characters: Consultation fee. So, why do we have to go to the workplace of the parties involved when doing research, and why have I never done an online video interview in so many years? Because the two are different. It's not a problem to interview familiar people through video, but when interviewing strangers, since there is no basic trust relationship between them and you, it's impossible to judge whether what they say is true or false. But if we go to the other party's workplace, we can see many things that cannot be seen through videos, which of course cannot be given to us by numbers.

Professor Su Li's article discusses the process of conducting research in legal sociology, which is filled with power relations and resources. As a researcher conducting field research, one must consider these questions: How did I come into contact with these people? What kind of power and resources do I need to use? Who made a phone call for me? And what impact will the relationship between the person who introduced me and me, as well as his relationship with my research subject, have on the trust and communication between me and my research subject? These are all very important issues. That is a very classic article by Teacher Su Li, which presents this question directly for the first time and explains it very thoroughly.

But today, more than twenty years later, with the rise of big data research, we are facing a new issue of power resources. So far, the domestic legal community has not reflected on this issue properly, but in the past two years, I have increasingly felt that it is a very important issue. After the rise of big data, the way we conduct research has undergone a change. This was most evident during the three-year pandemic, when we couldn't do interviews. We may be able to do them online, but if we want to go to the court for interviews, we can't enter without a nucleic acid test report, and having a nucleic acid test report doesn't necessarily mean we can enter. So in this situation, we rely more on the research method of sitting at home analyzing big data, and even in recent years, more and more new concepts have emerged, such as data law, digital law, artificial intelligence law, and even metaverse law... Personally, I am not optimistic about this trend. To quote a classic line from Jiang Wen's movie "Let the Bullets Fly": "You have to drink alcohol one by one, and take the road step by step. If you take too big a step, it's easy to pull an egg." Over the past few years, we have gone from data law to digital law to artificial intelligence law and then to metaverse law, but we have taken too big a step. Why are these new concepts related to big data emerging in such a short period of time? This is directly related to national macro policies and social development needs. At present, the development of artificial intelligence and various big data based network platforms in China is advancing rapidly, and some aspects have already surpassed European and American countries by several streets. From the perspective of national macro policies, it is necessary to encourage the development of big data and artificial intelligence. As a result, the legal community has been following suit, but this problem has arisen. Whether it is called data law or digital law, no matter how it is done, it highly relies on the data provided by the national system.

Twenty or even a decade ago, we could have conducted social science surveys without going through official channels. For example, my senior brother Ethan Michelson, who was studying at the University of Chicago, conducted two surveys on Chinese lawyers in 2000, one in Beijing and the other nationwide. He, a 'foreigner' and a doctoral student, actually conducted a city wide and nationwide questionnaire survey. It's very difficult for you to do the same thing now. Most of our current surveys or big data research require the support of the national system. If I want to conduct a questionnaire survey on lawyers in Yunnan Province now, it's unlikely that I wouldn't inform the Department of Justice and the Bar Association and do it myself. This is a very big problem faced by empirical research nowadays, which is that data must be attached to the formal system of the country in order to be obtained, which is very different from more than a decade ago. Whether it's when Professor Su Li wrote 'Sending Law to the Countryside: A Study of China's Grassroots Judicial System', when Professor Chen Baifeng wrote 'Rural Jianghu', or when I wrote 'The Logic of Secession: An Ecological Analysis of China's Legal Service Market', we don't need too many resources within the system, just do our own research.

I didn't spend much money on field research for my doctoral thesis. I went to 12 provinces and conducted over 200 interviews, spending only over 50000 yuan in total. Of course, that was the price from 2006 to 2007, and even if converted to current prices, it would only be a few hundred thousand yuan at most. But now any big data research costs millions or even tens of millions of yuan, and I don't know how this money is spent. Field research in social sciences does not require spending so much money, but now doing so-called data law and digital law based on big data wastes a lot of national resources, and makes legal researchers, whether teachers or students, rely entirely on materials and data from within the system for their research. This is a very big problem. Since these data can be provided to you, they can also be taken away, just like a faucet. China Judgments Online is a good example. The faucet was turned on as soon as it was said, and the data on the internet kept flowing. But now suddenly it is said that the faucet needs to be turned off, and it may not be available next year, leaving everyone without rice to cook. What about us scholars and students who rush to engage in data law? You see, I don't have any pressure at all because I haven't taken this step yet.

This is a deeper ethical issue in research. As scholars, to what extent should we maintain our academic independence and not let academic research become a tool completely dependent on the system. Regardless of our political stance, it is crucial to maintain the independence of scholars in academic pursuits. I am studying professional sociology. If a scholar lacks independence in academic research, then their professional autonomy will be lost. Without autonomy, their professional status will not be too high. This is a deeper issue triggered by the "scientific turn" of social science and law in the past decade.

3、 How to Break the Superstition of Numbers (Quantitative, Data)

Having talked so much, it seems like I am criticizing others. So, how can social science and law research "break the superstition of numbers" and do more research that is not so "pseudoscientific" in the future? I mainly want to talk about two issues, one is how to do better at the methodological level, and the other is how to tell the story better in narrative.

To "break the superstition of numbers" from a methodological perspective, one must first realize that there are misconceptions in both the Chinese and American legal circles. My personal opinion is that the methodology of legal empirical research is inclusive, and both quantitative and qualitative methods are legal empirical research methods. But some people in the American legal community have different views. The English term for empirical legal research is "empirical legal studies," but what they define as "empirical legal studies" is actually quantitative legal research based on statistical analysis. Studies in legal anthropology or legal sociology that use qualitative methods like mine are not considered 'empirical legal studies'. There is still a good journal called Journal of Empirical Legal Studies, but if I write an article using qualitative data such as interviews and submit it to this journal, the article will probably not be sent for review and will be rejected directly. They are methodological supremacists, which shows that this problem is not limited to China. However, this tendency reflects a very erroneous assumption that quantitative methods are more scientific than qualitative methods, and that research conducted through questionnaire surveys and regression analysis is more scientific than research conducted through qualitative methods such as interviews and participant observations. They even have a pseudo scientific arrogance, thinking that what I have is advanced and what you have is backward; By using these statistical methods, I have become "sophisticated" and more advanced than you. If I can do statistics, I can study anything, while those of you who do qualitative research use more "primitive" methods because you don't know how to do statistics.

But in reality, it's not like that. Nowadays, the country's network technology, artificial intelligence, and social media are very advanced. The progress of these technologies not only brings big data and makes quantitative research easier, but also makes qualitative research easier, and new methods and possibilities have emerged. Give two examples.

The first example is my student Li Sitao at the University of Toronto. He is very interested in grassroots criminal justice in our country, but because he is pursuing a doctoral degree, he has been abroad for the past few years. Although he has returned to China, he cannot enter the court during the epidemic and cannot participate in observation or attend court hearings. But Li Sitao found a way, because he is from Changchun, Jilin. He found that in the past few years when judicial transparency was being carried out, many trial videos using simplified procedures in the grassroots courts of Changchun would be online and could be seen on the court website. He finds this very interesting. Although he cannot enter the court, even if he is abroad, thousands or tens of thousands of kilometers away, he can still watch this video. Of course, watching videos themselves is not perfect and cannot replace being in person. It's like why everyone is listening to lectures in this room, even if the seats are full, so many students still sit on the ground and stand at the back of the classroom, because the interaction between people in the same room can never be replaced by watching videos. But in this situation where one cannot enter the court, watching videos becomes a very good alternative.

Li Sitao published an article in Law&Society Review, which is the top publication in the field of legal sociology worldwide. This article is called "Face Work in Chinese Routine Criminal Trials", which translates to "Face Problems in Chinese Daily Criminal Trials". How did he study the issue of judges' face? He selected over 100 online videos from a grassroots court in Changchun, watched them from beginning to end more than once, and even made records, spending a lot of time. Then he observed the interaction process between the judge and the parties, the prosecutor, and the defense lawyer from the video, and saw some particularly interesting phenomena. Li Sitao found that judges have both "face" and "face" in court, with two types of faces: "legal face" and "political face". What is a 'legal face'? It means that the judge should adopt a legal posture during the trial. Although many criminal trials follow a procedural approach, the defendant has already confessed and accepted punishment, and the outcome is likely to be similar. However, judges still need to maintain this "legal face" by going through the process from beginning to end, with specific ways and tones of speaking. These are important ways for judges to uphold judicial authority. But sometimes the "legal face" works, sometimes it doesn't. Some defendants may challenge the authority of the judge and expose the judge's "legal face". At this time, the judge has to put on a "political face" and give the defendant preaching political education, which is also common in court trials. In addition, he also found that judges have a "face" issue, and an important point during the trial process is that all participants intentionally or unintentionally maintain the judge's face. For example, if a judge sometimes says the wrong thing or even makes a mistake about the case, prosecutors and lawyers will find ways to help the judge maintain his face, which is also an interesting discovery.

I won't go into too much detail about the content of this article. You can think about how this research was conducted, as it was entirely done through high technology. The emergence of videos is a very recent thing, and high technology has brought us not only big data, but also a lot of other types of data. For example, observing court hearings, there are now countless times more materials than before. Compared to on-site observation, video has its drawbacks but also its benefits - you can watch a video repeatedly, such as the example I just gave, where the judge said the wrong thing and the prosecutor and lawyer reacted. If there is a video, you can watch it 5 or 10 times and see many subtle things, but if you watch it on site, you may overlook them. After reading it 5 times, you will find that a certain sentence seems to be implicit, because in fact, many people within the system, including judges and prosecutors, speak very fluently. Sometimes they want to tell you something, but they cannot say it clearly, so they will tell you in a very gentle and technical way. This kind of thing may pass away after listening to it in court, but watching that video 10 times may reveal it. So video has its benefits, it allows us to observe many things in depth and detail. Originally, we could also participate in observation in the court by taking paper notes. However, when you take paper notes or record audio, many things cannot be recorded, such as people's expressions. Of course, the clarity of our court trial videos is not yet that high, to the point where we can clearly see the judge's face and facial expressions, but we can still see or hear him smiling, angry, and his tone rising. These are not numbers or statistical methods, but they are also new possibilities brought to our empirical research by high technology.

So when conducting research and writing articles, you can consider what new research methods and data sources are available now, such as the video observation method used by Li Sitao, which is a new method that can be used. Will the quality of the research made using these things to stir fry dishes be slightly higher than the quality of using tens of thousands or millions of judicial judgments to "draw a cake to satisfy hunger" or "cook white rice"? Alternatively, the two can be combined, with both rice and stir fried dishes, and this research may be more comprehensive.

The second example is about online ethnography, which is also a new possibility brought about by technological development for field research in legal sociology. A few years ago, I wrote two articles with another student, Wang Di, on this issue. Traditional ethnography is done offline and requires going to the "fields" to participate in observation, such as working in a court for a period of time and participating in the daily lives of judges; If you study lawyers, you have to work at a law firm. But in today's social life, everyone's daily life cannot do without the internet. During the three-year epidemic period, over 90% of our social activities may have been completed online. Even if the restrictions on epidemic prevention are lifted now, we may spend more time interacting with our classmates, friends, relatives, and parents online than offline every day. This kind of social interaction on the internet has become a very important part of our daily lives, so as researchers, whether studying lawyers, judges, or citizens' legal awareness, it is impossible to be separated from the internet. This has given rise to a new research approach: online ethnography.

There are many similarities between online ethnography and offline ethnography, but there are also some obvious differences. Traditional ethnography involves conducting field research in a place for a period of time, and then returning home, possibly never going to that place again and never seeing these research subjects again. As researchers, one can enter and exit the fields, and it is not difficult for traditional legal anthropologists to isolate themselves from them after leaving the fields. I interviewed over 200 lawyers while conducting my doctoral thesis research, but my contact with them decreased afterwards. Some lawyers may later become friends and occasionally have contact, but this situation is actually rare. The vast majority of lawyers have not contacted them since I interviewed them once. However, this is already quite difficult now. I need to interview a lawyer, and the first step is to add WeChat. Not adding WeChat is unlikely. Adding WeChat means he will always be your friend, unless he is blocked or blacklisted. So, as a researcher of online ethnography, once you enter, you cannot leave.

For example, if you want to study a certain legal profession group, they will add you to a WeChat group where you have met many judges, lawyers, or law students. You interact with them in the group and even interview some people offline. You see a lot of things in the group that can become data and materials for your research. After the research is completed, you can of course leave the group, but as long as you leave the group, they won't be able to find you anymore? Especially after implementing the real name registration system on the internet, it is highly unlikely that you will disappear from this world. So this comes back to the power resource issue that Mr. Su Li said, because our Chinese custom is to mutually benefit each other. I ask you to do things today, and you will ask me to do things later. Maybe I begged you today. You won't find me for 10 years, but you may find me to do something after 10 years, which is the human nature of our Chinese people. But now, when you enter the field, they provide you with a lot of things. After you come out, one day, for example, a interviewee comes to me and says, 'Teacher Liu, I have something I want to ask for your help with. What should I do?'? This is a very practical problem.

Online ethnography is also a huge challenge in ethical research. There is currently no academic ethics review system for social science research in mainland China. This system exists in many countries and regions, including the Hong Kong Special Administrative Region of China. The survey I am conducting on the living conditions of young lawyers also took me over a month to go through the academic ethics review process at the University of Hong Kong, because as a social science researcher, it is the bottom line to ensure that our research does not cause harm to the research subjects. This thing is easy to say, but traditional academic ethics review has a series of very rigid regulations, such as the principle that the interviewee must be anonymous, and I definitely cannot disclose their name or affiliation information. So, all the interviews in the book 'The Logic of Separatism: Ecological Analysis of China's Legal Service Market' were only written about provinces. For example, the interviewee is a lawyer from Heilongjiang Province, and I don't even know which city they are in because I want to fully protect their personal information. In addition, obtaining the consent of the interviewee is also necessary for conducting interviews. In theory, the most standard form of consent in foreign countries is to provide the interviewee with a piece of paper, an informed consent form, and even a signature for each interview. This method is not very feasible in China. For example, if I want to interview a lawyer and I come up with a piece of paper and say that this is my ethical rule for the interview, and you sign it for me, the interviewee will have been scared away long ago. So, although you don't necessarily have to sign a written document, at least you need to inform and obtain consent. You can't just go for an interview or participate in observation just because someone doesn't agree. But online ethnography presents a new challenge, such as the example I just gave. If someone adds me to a WeChat group, I don't know all the people in the group, and some even use usernames. I don't even know who they are, how can I get their consent? If consent cannot be obtained, can the things I observed in the WeChat group be used as data? According to the traditional rigid standards of social science academic ethics review, this data material may not be usable because your research subjects have not agreed.

This is the new problem that the rise of online ethnography has brought to our empirical research, but at the same time, it has also brought many new opportunities. I am reflecting on this issue now and becoming increasingly convinced of it. Our future ethnography will mostly be a combination of online and offline, and pure offline ethnography is likely to become increasingly rare. Because as I just said, a large part of human interaction is now done online, not just between researchers and research subjects, but also between ordinary people. This is also a new possibility provided by technology for social science research, especially qualitative research.

In short, we need to break the "digital superstition". It's not that breaking the digital superstition means not conducting empirical research, but rather that breaking the "digital superstition" can help us see through the clouds. After removing these "dark clouds" of big data, we can see that there are actually many better, newer, and clearer possibilities for legal empirical research. New technologies, data, artificial intelligence, and even the metaverse may bring us new fields and new application spaces for our interviews, participatory observations, and other qualitative research methods.

Finally, let's talk about the issue of storytelling. Why is social science and law becoming more and more like 'pseudoscience'? One important reason is that articles nowadays are becoming less storytelling, which is very different from a decade ago. Why is Chen Baifeng's' Rural Jianghu 'a very good book? There are many reasons. He summarized four principles from Mr. Fei Xiaotong's understanding of Chinese rural society in his book "Rural China", and then developed a theoretical framework from these four principles in his book. He used this theoretical framework to analyze the "hooligans" problem in the Two Lakes Plain, which was done very well in theory. But there is one more important point, Professor Chen Baifeng came from the Huazhong Village Governance School, and the two most important words in their research on Huazhong Village Governance are "deep description". What is deep drawing? You must tell the story very carefully and tell it well. Even in theory, there is no need to write fancy stories, but the story must be told well, using stories to persuade people and use stories to move people. This "deep description" writing style is very important in social sciences and is more commonly used in history, where historians are all telling stories. In fact, whether it is historians or sociologists, even for those who conduct quantitative research, the biggest difference between first-class scholars and second-class scholars is that first-class scholars can tell good stories. Some second rate scholars have excellent statistical analysis techniques and complex analysis results, but they cannot tell good stories, so their articles cannot be written well, which is a huge gap.

I met a historian in a meeting many years ago who made a particularly good analogy. She said that for historians, theory is like underwear, you should always wear it, but you should never reveal it. Historians may not have much theoretical knowledge when it comes to writing, but it's not that they lack theoretical foundation. Their theoretical foundation is much deeper than ours, the 'pseudo scientists'. But they won't talk about a big theory, because theory is like underwear, hidden in words, and all you see is the clothes on the outside. So after listening to that sentence, I had a sudden realization that I had been studying sociology for so many years, and it turned out that I only wore underwear on my head all day! What we pseudo scientists are basically doing is this. But then again, how can we tell the story well without being so 'pseudoscientific'? This is very knowledgeable. My supervisor Professor Abbott at the University of Chicago wrote a book on methodology called "Methods of Discovery", which teaches people how to write articles and narrate stories. He believes that there are basically three narrative styles in social sciences: the first is causal, the second is analytical, and the third is narrative. The narrative style is actually very similar to the deep description style used by the Huazhong Village Governance School, which is to tell a story. There is no need to pull up so many big theories or put underwear on the head, just tell the story well.

French sociologist Bruno Latour once said, "Only bad descriptions need explanation, good descriptions do not need explanation." I do not completely agree with his statement, as explanation is still important for sociology, but there is a profound truth in his words. Writing the article in a deep and descriptive way, if it can really be written well, is actually more advanced than the causal and analytical writing styles. Because the story is told clearly, there is no need to talk about big theories anymore. In fact, the work of description and explanation is done together, and this kind of research is definitely more touching. This is also the most lacking aspect of current social science and law. A few years ago, during the epidemic, I wrote an article for the China Law Review, which was about the three narrative methods of legal sociology, taking the COVID-19 as an example. I call these three methods numbers, systems and people's hearts. The most lacking aspect of current research is the human heart. In Weber's "Protestant Ethics and the Spirit of Capitalism," there is a famous saying that goes, "Experts have no soul, and those who indulge in pleasures have no heart." Many current social science and legal studies have neither soul nor heart. Social science research is ultimately about people, written by people and written for people to read. If a study cooks a pot of white rice from beginning to end, full of big data, but without human emotions or emotions, then this study is definitely not the best. The best research, whether using quantitative or qualitative methods, should result in articles that are compelling and moving.