[Special Feature: Yukichi Fukuzawa and Statistics] Yukichi Fukuzawa's Statistical Thinking / Keita Yamauchi

2020/06/05

In recent years, in what is called the era of big data, interest in statistics and statistical education—from schools to working adults—has been rising. In the November 2014 issue of Mita-hyoron (official monthly journal published by Keio University Press), the special feature was "Statistics Opens the Future." Furthermore, coupled with expectations for artificial intelligence (AI), the ability to utilize data, discover problems, and solve them is increasingly being demanded.

In this context, April 2020 marked a major turning point. Under the new Courses of Study, statistical education from elementary to high school was restructured, and new textbooks based on these guidelines began to be used simultaneously in elementary schools nationwide this spring. As a member of the Minato City Board of Education, I compared textbooks from all publishers for all grades to select those for use in municipal elementary schools, and I truly felt the shift of the times. Furthermore, at Keio Yokohama Elementary School, for which I was responsible for establishing, "education in the power of language," including the language of numbers, is one of the three pillars as the foundation of logical thinking. In my graduate research and education in medical management, I have also aimed for the application of exploratory and interactive data analysis. For these reasons, I am deeply moved that such an era has finally arrived.

For those who received a traditional education, memories of statistical education might center on practicing calculation methods for averages and probabilities. However, in the mathematics curriculum of the new Courses of Study, not only has the volume increased, but the content has shifted to emphasize "statistical problem-solving." It advocates learning these methods to "reach one's own conclusions regarding problems and critically reflect on the validity of those conclusions" and "acquire an attitude of utilizing this in life and learning." In fact, the new textbooks are designed to select subjects that facilitate discussions for problem-solving. Even more symbolic is the inclusion of the PPDAC cycle, also known as the "statistical inquiry process." This is a way of thinking that continuously rotates through the cycle of Problem (clarifying familiar issues) → Plan (considering data to collect and how to collect it) → Data (collecting data) → Analysis (creating tables and graphs to analyze and find patterns) → Conclusion (presenting a conclusion to the initial problem and identifying new issues). As this cycle shows, there is great significance in the fact that the emphasis has shifted from merely acquiring statistical techniques to education in statistical thinking, or the thinking process of scientific problem-solving. In fact, this thought process has been emphasized in statistical education worldwide as the core of 21st-century skills.

While the relationship between Yukichi Fukuzawa and statistics has been discussed before, here I would like to consider it specifically from the perspective of statistical thinking.

The first person to point out that Yukichi Fukuzawa initiated Japanese statistical history and played an important role was Masao Yokoyama, who studied statistics under Koji Sugi and succeeded Ayatoshi Kure to lecture on statistics at Keio University. Yokoyama noted that the first translated publication in Japan to use the term "seihyo" for "statistik" was "Bankoku Seihyo" (translated by Setsuzo Okamoto [later Masao Furukawa] and published with Fukuzawa's proofreading, a collection of statistical data from countries around the world). He also noted that many disciples, including Tokujirō Obata, participated in the founding of the Statistical Association; that disciples including Fumio Yano and Takuzo Ushiba participated in the Statistics Board established in the government at the suggestion of Shigenobu Okuma; and that Meiji Life Insurance, founded by Taizo Abe, was the pioneer of the life insurance industry, which is one application of statistics. Research into such achievements has become increasingly sophisticated in recent years.

However, it is necessary to note that Yukichi Fukuzawa did not merely import or create statistical materials; he was already advocating a way of thinking that leads to the statistical thinking mentioned at the beginning.

For example, there is a passage explaining "jitsugaku (science)" in the first volume of Gakumon no susume (An Encouragement of Learning). It is the part containing the famous phrase, "What should be pursued with all one's heart is jitsugaku (science) that is close to the ordinary daily use of humans." While this single phrase is sometimes taken out of context, leading to the misunderstanding that "jitsugaku (science)" is merely a practical or useful study, if one reads further, it says:

"In every branch of learning, one should grasp the actual facts, follow the matter, and seek the reason of things nearby to meet the needs of today."

In other words, he points out that in any field of study, one should first observe facts and phenomena, pursue the theories and laws behind the events based on those objective facts, and apply them to daily life—that is what he calls "ordinary jitsugaku (science) for humans."

Incidentally, the 12th volume of Gakumon no susume (An Encouragement of Learning) also contains a frequently quoted passage regarding "jitsugaku (science)": "The essence of learning lies only in its application. Learning without application is equal to ignorance." Following that is this passage:

"The main purpose of learning is not just reading books but the work of the spirit. To utilize this work and apply it to actual situations, various efforts must be made. 'Observation' means inspecting things. 'Reasoning' means investigating the logic of things and attaching one's own theory."

Here, too, the importance of "observing" phenomena and "investigating" their theoretical laws is written. Furthermore, he points out that this alone is insufficient, and that it is necessary to "exchange" that "knowledge" through conversation and "disseminate" it through books and public speaking. Incidentally, while "inquiry" (tankyu) is a keyword in the new Courses of Study, Yukichi Fukuzawa sometimes used the word "investigation" (suikyu).

In this way, Yukichi Fukuzawa used the term "jitsugaku (science)" with the meaning of a scientific way of thinking. For example, in the "Keio University Records," which could be called the 25-year history of the school, it states, "The principle of this school is contrary to the old school of Japanese and Chinese learning; even when discussing literature or talking about worldly affairs, it is based on Western jitsugaku (science)," and the word "science" is written as a gloss over "jitsugaku."

While writing works for adults, Yukichi Fukuzawa also energetically wrote books for children of what we would today call elementary and junior high school age. These can help us understand the ideas he valued. Because he strove to make things important for the new era easy to understand for children nationwide, his thoughts are expressed frankly. For example, there is a calligraphy textbook called "Keimo Tenarai no Fumi." While previous textbooks used Chinese poems, waka, or letters for ceremonial occasions as examples, he used new views of learning as examples. Among them is the following sentence:

"The purpose of physics is to fix one's eyes on things that people are usually accustomed to and do not find strange, to observe things that people do find strange and search for their logic, and to apply this to practical use to benefit the world; this is the first and foremost duty."

Does this not also show the thought process of scientific inquiry in a very easy-to-understand way?

The representative work in which Yukichi Fukuzawa mentioned "statistics" not as statistical data but as a discipline for investigating tendencies and laws in social phenomena is An Outline of a Theory of Civilization.

Yukichi Fukuzawa wrote this after thoroughly reading Buckle's "History of Civilization in England." Masao Maruyama introduces this book in "Reading 'An Outline of a Theory of Civilization'" as follows:

"The shorter the time and the narrower the space, the greater the dominance of chance. Room for individual special circumstances emerges. Through mass observation, special circumstances are offset, and general laws emerge. The method of mass observation that Buckle cites is (...) statistics."

And he evaluates An Outline of a Theory of Civilization this way:

"This 'Outline' by Fukuzawa, who learned from Buckle, was the first to grasp this (statistics) as a general method for objective recognition of social laws, independent of policy theory."

Based on Buckle's book, Yukichi Fukuzawa introduced the fact that there are certain tendencies in the annual number of murderers, suicides, and daily sales of steamed cakes. He further showed that by tabulating and comparing population, prices, wages, number of marriages, births, illnesses, and deaths, the situation of a country becomes clear at a glance, and that the annual number of marriages is related to grain prices. Based on this, he stated that investigating circumstances in this way is very convenient for seeking causes, and that it is necessary to consider that causes include "proximate causes" and "remote causes." Incidentally, comparison and inference of causal relationships are important concepts in statistical thinking.

This part is based on Buckle's book, but the descriptions of the examples given feel as if he has truly made them his own. He began writing this book after refining the concept around February 1874. Considering that the aforementioned first volume of Gakumon no susume (An Encouragement of Learning) was published in February 1872 and "Keimo Tenarai no Fumi" in 1871, for Yukichi Fukuzawa, who had already recognized and discussed the importance of scientific inquiry thinking, the statistics shown by Buckle must have been both attractive and understandable with a sense of reality.

He also criticized making groundless assumptions by being obsessed with one thing, saying "the state of the world should not be judged based on a single matter or object," and criticized failing to seek remote causes by being "infatuated with what is immediately heard and seen," explaining the significance of statistics. Since groundless assumptions and infatuation were the thinking attitudes Yukichi Fukuzawa most rejected, he likely strongly agreed with statistics as a way of thinking to avoid falling into them.

The same book also contains the passage: "Even if one tries and makes good progress, I have not heard of anyone reaching the limit. From the beginning of creation to the present day, it may be called a world of trial." This resonates with the thinking of continuously rotating the cycle of statistical inquiry.

Yukichi Fukuzawa's understanding of statistics and the statistical thinking symbolized by "jitsugaku (science)" were significant in that they were demonstrated even where concrete data was absent, but I would also like to look at how he handled numerical data.

Yukichi Fukuzawa was originally a person who actively collected numerical information and tried to understand it concretely. For example, in his notebook from the 1862 mission to Europe, where he explored the situations of various countries, many numbers are written down vertically and horizontally.

In his writings, he not only used statistical materials to explain a country's situation more concretely and clearly but also frequently showed comparisons to advance his arguments. Looking at the editorials of Jiji Shinpo, in "Meat Must Be Eaten" (December 16, 1882), he estimated the per capita beef consumption from the number of cattle slaughtered across Japan, and further added generous estimates for pigs, boars, sheep, etc., for which there were no figures, to show that Japan's consumption was significantly lower compared to Western countries. In "On National Wealth" (June 23, 1883), he tabulated the volume of various types of sake between 1873 and 1882, demonstrating that the demand for refined sake temporarily decreased due to tax increases but returned to previous levels the following year.

In this way, it can be seen that he daily employed thinking such as using statistical data to make comparisons, observe changes, or consider factors.

As a concrete example, let's look at Transition of People's Way of Thinking (published in 1879). This argues that "recent civilization" (also written as modern civilization) of the 1800s is a civilization of steam, and that the feelings of people in countries around the world—the "people's spirit"—have been "renewed" by the "conveniences of civilization" such as steamships, trains, telegraphs, printing, and mail.

In this, he explains that these tools have brought about an era where "information" (also written as information) spreads instantly. After showing the circulation of newspapers that "broaden people's information" in the UK, he shows that the number of postal letters that "exchange people's information" was over 780 million in 1867, flourishing at 25 letters per person, and increased to 967 million in 1874.

Around this time, Yukichi Fukuzawa seems to have viewed the situation in Japan in contrast based on statistical materials as an essential element for the progress of civilization. In "Popular Theory of Civil Rights," published the previous year, he wrote, "To measure the degree of progress of human intelligence, one can know a part of it by looking at the amount of mail in that region," and listed the number of pieces of mail for one year starting from 1876. Specifically, he listed Tokyo Head Office at 15,103,000, Osaka Office at 2,788,000, Kyoto Office at 1,461,000, Aichi at 965,000, and Aomori at 257,000. He further divided Aichi and Aomori by population to show that they were low at 0.8 and 0.55 letters per person per year, respectively.

In Transition of People's Way of Thinking, he further points out that people in an era when information transmission was slow were like pupae in terms of butterflies, and the government could easily control them. However, people who have obtained the "tools for communicating thoughts" are like butterflies with wings and cannot be easily controlled. He notes that some countries take measures of despotic oppression, but such old-fashioned ways will not succeed. He introduces the British form of "transferring political power during peaceful times" and "constantly alternating political power between old and new," and states the necessity of a national assembly.

Here, too, after publishing a list of the "date of appointment," "term of office," and "name of the ruler (Prime Minister)" for Prime Ministers between 1784 and 1879, he showed that "during those 96 years, there were 26 generations of changes in rulers; the shortest term was 121 days, the longest was 17 years and 84 days, those over 5 years were (...) 7 people, and those over 10 years were only 2 people. Also, if these 96 years are averaged over 26 generations, the term of office for one generation is a little over 3.69 years," and "one can see that the turnover is rather fast." It is also noteworthy that he carefully recorded not only the average but also the distribution of the terms of office, including the maximum and minimum values. Additionally, it is interesting that he also created a table for the years in office of senior councillors and others in the Tokugawa government in Japan, showing the average and distribution to explain that it was not a position held by one person for many years. In this way, it is clear that Yukichi Fukuzawa had the thinking to actively collect data himself, look at distributions, and make comparisons.

This kind of thinking did not decline until his later years. For example, the Jiji Shinpo editorial "The Necessity of Statistics in Religion" (April 24, 1898) argued for the necessity of reform, stating that the situation of Buddhism was marked by problems with the "conduct of the head priests" and there was no hope of maintaining it as it was. He then asserted, "My plan regarding that reform is to judge the diligence or laziness and ability or inability of monks based on statistical figures when performing their promotion or demotion." He proposed that monks should be evaluated by looking at the number of households and temples, the number and ratio of population and monks, the time spent on sermons by monks, and the number of police crimes in a year. The proposal to evaluate monks based on structure (population ratio of monks), process (sermon time), and outcome (number of crimes)—three elements often used today in evaluating the quality of services in hospitals and other institutions—must have been novel for that time.

After An Outline of a Theory of Civilization, Yukichi Fukuzawa did not often discuss statistical thinking directly. However, those thoughts were expressed in One Hundred Discourses of Fukuzawa, published in 1896.

In Discourse 86, Fukuzawa went so far as to say, "Those who lack this overall statistical thought are not worthy of discussing matters of civilization." In Discourse 34, he stated that "what I have advocated for many years is the jitsugaku (science) of civilization," and said that the achievements in science and technology, including steam and electricity, resulting from the "jitsugaku (science) of investigating various subjects"—which stems from the characteristic of civilization to "calculate the reality of mathematical principles and dissect the large and small" and "raise doubts to investigate the source"—are countless. He then said that even looking at intangible areas apart from science and technology, namely politics, law, and economics, "their progress and development are nothing if not the gift of mathematical principles. The fact that people in Western countries have long valued the methods of statistics and utilize the actual numbers of statistics to observe all human movements to seek the greatest happiness for the greatest number is also sufficient to glimpse where their thoughts lie." Finally, I would like to introduce the following passage:

"When explaining the reason why those trains and steamships move based on academic principles, showing the principles of statistics through the management of mail, and talking about the wonderful use of electricity by proving it with the actual state of telegraphs, one sees nothing that lies outside of academic principles in all things. Not limited to the new things of the civilized style, even the smallest tree or stone, or a single piece of paper or a hair before one's eyes, if one explains its nature and clarifies its utility in light of the principles of truth, and gradually investigates its logic to enter the profound and reach the most profound of the profound, the universe will be encompassed within the human mind (...)" (Discourse 70)

Seventeen years after analyzing the postal situation in Transition of People's Way of Thinking, the 62-year-old Yukichi Fukuzawa held expectations that, just as there are theories in science and technology, social phenomena and utilities could also be investigated through statistical thinking, expanding its scope.