The Skew in the NDC Categories Just After the End of the War

The NDC is a way of classifying books by a combination of three digits — that thing on the sticker on the spine of a library book.

I compared the NDC class shares of 3,564 titles published just after the end of the war (June 1945 to September 1946), taken from the National Diet Library Digital Collections (the Prange Collection excluded), against 1,580,026 titles covering all periods.

I divide the share just after the end of the war by the overall share to get a "ratio." Above 1.0 means relatively more of it just after the end of the war; below means less. This shows what got published a lot in that period and what did not.

Right now I am reading every preface and afterword of 6,037 titles from this period. Halfway through, I wanted to check the feel I had picked up against numbers, and so I tried this analysis.

Method

1. Extract the data from the search results using AI
2. Analyze the dataset with AI
3. Go over the results closely and match them against my own findings

The Mood at the Time

Judging the past by numbers alone leads you astray. So let me share the mood at the time first.

To begin with, there was a climate of opinion that things useful for establishing the new country ought to be published first.

Some may scold me, saying that this is no season for idling away at fishing and the like, but that is an outrageous misapprehension.

*Fishing Through the Four Seasons* (Shiki no tsuri), Masuda Hajime, Yūbunsha, 1946.https://dl.ndl.go.jp/pid/1125914/1/7

The study and practice of this must be nothing less than the urgent task of us who live under the present situation.

*What Is a Citizen?: An English-Japanese Parallel Text* (Kōmin to wa: Eiwa taiyaku), by Fujimori Seiichirō, Kōmin Kyōhonsha, 1946.https://dl.ndl.go.jp/pid/11177473/1/13

Why was there an order of priority? Because the printing houses had been hit.

The many illustrations that ought to have been included in this volume were lost with their plates when the printing house was destroyed by fire, and we have had no choice but to omit them. They may be supplied on another day, at the time of a second edition.

*An Outline of Chinese Literature, Part One* (Shina bungaku gairon, jōhen), Shionoya On, 1878-1962, Kōdōkan, 1946.https://dl.ndl.go.jp/pid/1150812/1/6

The writing of this book began in the early summer of 1944, was broken off any number of times, and came at last to a provisional completion in February of last year.

But with the publisher and the printing house burned out and the proofreaders scattered, the author had all but given up on publishing it. That the author has been able to bring this book into the world while still alive is closer to a miracle. If this book, born under such a fate, can contribute even a little to scholarship, that would be a happy thing.

*Fiscal Statistics* (Zaisei tōkei), Shiomi Saburō, 1895-1962, Daiichi Shuppan, 1946.

Complicated printing that included figures was evidently difficult too, and many hand-written mathematics study guides survive.

Democracy and elections, the abnormal expectations placed on a nation built on science, the impulse to entrust the future to children, the fact that people of the social standing to write books are not as gloomy as you would expect — keep all of that in a corner of your mind as well.

Share by Main Class (1-digit NDC)

Literature (ratio 2.06) and natural sciences (ratio 1.67) stand out.

History (0.54), industry (0.61), and the arts (0.61) sink.

Classes Overrepresented Just After the End of the War (2-digit NDC, Highest Ratios)

Classes 97–99 have no corresponding code on the overall side because of a difference in the edition of the data, so the apparent outliers there have been excluded (see "Notes on the Data" below).

Classes Underrepresented Just After the End of the War (2-digit NDC, Lowest Ratios)

Inside the Major Fields

The Sciences: A Concentration on "Practical Learning"

In mathematics, physics, and chemistry alike, the books are not basic theory or systematic scholarship but lean toward practical areas tied directly to reconstruction — calculation techniques, experimental methods, and the like.

Meanwhile, fields where the direct application is harder to see, such as algebra (411, ratio 0.55), are flat.

The defeat was understood as "the defeat of science and technology," and you can feel the rush to get practical science and engineering education going. I also keep running into drill books for the abacus. Short, practical, easy to print — that may have been the reason.

Japanese Literature (Class 91): The Weight of Poetry and Essays

Of the 753 titles in Japanese literature, the breakdown runs: novels and stories 342 titles, poetry 208, criticism and essays 105.

Criticism and essays have the highest ratio, 4.36. This seems to have had a lot to do with the mood of the times.

Up to now it has been chiefly the long work that found favor in Japan; but judging from the actual state of the world, I have the feeling that henceforth the short piece will be the welcome one. What I mean is that, with such great national and popular undertakings as the economic reconstruction after the defeat, the building of a new Japan, and the movement for democratization all whirling about us in a dazzling storm, the people seem unlikely to have the leisure to sit down quietly and savor a long work.

There must also have been the physical constraint that, with materials short, only books with few pages could be brought out.

The ratio of 0.16 for records and reportage (916) stands out as extremely low.

• A movement to record and examine the experience of the war objectively was slow to appear just after the end of the war
• Records of war damage had a hard time getting past the censors

Such factors can be imagined.

As an aside, a fair number of literary men were insensitive and unreflective.

The remaining ten percent, that is, everyone but the ninety percent of the nation, had promptly confirmed the "fact" of the "defeat" and had correctly foreseen "what would come after." They did not fall into any unseemly panic. They knew that "what was to come" had arrived by law. They thought it not in the least strange that the "divine wind" had failed to blow. And who were that ten percent?

Prime Minister Shidehara may perhaps have been one of them.

But… that ten percent was no one else than the men of culture. And among the men of culture, the literary men above all. The "bandwagon riders" and the "tea-serving flatterers" are of course out of the question, but at the least, the literary man as a man of the study. Then the scientists. Then a very small number of religious men. Were these not the citizens included in that ten percent?

I think on it. To state the reason in a word: they have always been "apostles of truth," and have no wish to bow their heads to kings or to power. Not only that, but the man of the study, as the characters themselves show, makes his business of reading. He would have been versed in Japanese history, and at the same time he had no small interest in the cultural history of the world and the history of mankind.

*Talks on Life* (Jinsei dangi), Takeno Tōsuke, 1899-1966, Yōyōsha, 1946.

https://dl.ndl.go.jp/pid/1135497/1/4

Beyond the fact that the competition over what position to take had already begun, Takeno was a fair eccentric himself; he may have been an "apostle of truth" with "no wish to bow his head to kings or to power," but from an ordinary point of view this is a dreadful thing to say.

Mushanokōji Saneatsu and Shiga Naoya, too, put out collected-works sorts of volumes without a single word about their own conduct during the war.

This is a hypothesis, but tanka and waka had societies all over the country, so the damage to them was slight, they could get support from wealthy local families, and that may be part of why they were able to publish.

Translations of Western Literature: Russia, France, Germany

※ No code on the overall side (see the notes on the data below)

The presence of the 47 titles of Russian and Soviet literature is conspicuous. Under the occupation, interest in the Soviet Union was rising. There was also a fresh appraisal of Tolstoy, who had been popular to begin with. On top of that there were hopes pinned on communism. Books praising China and Russia stand out as well.

Social Sciences (Class 3): Building a New Order vs. Sluggish Practical Work

While the fields bearing on the foundations of a new order — the form of the state, the constitution, international law, social thought — increase, practical books on business management and commerce collapse.

Education (ratio 0.41) probably reflects both that a fundamental rewriting of textbooks was under way, which made new books hard to issue, and that the education given during the war had been plainly wrong.

Agriculture: Only the Food-Related Subjects Stand Out

The ratio for agriculture as a whole (61) is a somewhat low 0.87, but food crops (616, ratio 3.45) and vegetable gardening (626, ratio 5.24) alone stand out. Practical books on "what to plant and how, so that you can eat" got published.

Seismology (453): The Outlier at Ratio 19.59

Within the earth sciences, seismology alone shows an abnormal ratio. 『It is related to the fact that this was a period of consecutive great earthquakes — the Tōnankai earthquake of December 1944 and the Mikawa earthquake of January 1945』 is the AI's analysis; what actually happened is that the Earthquake Research Institute of Tokyo Imperial University kept placidly issuing the "Bulletin of the Earthquake Research Institute, Tokyo Imperial University," which makes the apparent number abnormal. It is a freak event produced by the small total number of publications.

The Shape of the Fields That Sank

The fields tied to the ideology of the national polity — Japanese history, Shinto, Buddhism, national defense — all shrink at once. This is likely the result not only of GHQ's publication restrictions but of self-censorship on the part of authors and publishers combined with a shift in what readers cared about.

As for sports, on top of having no margin to spare, the shortage of food must have played a part. Since before the war sports had been tied to improving the physique of the nation, so the mood was probably: no food, and you want to talk about physiques?

NDC Classes That Look Like Study Guides and Textbooks (Just After the War, 1945-46)

It is only an impression, but I keep running into study guides. So out of the 3,564 titles in the whole set, I pulled out the classes that correspond to practical science-side textbooks and study guides.

Mathematics (Class 41): 81 titles, ratio 2.77

The Prange Collection is not included, so the counts are small and the data is only good for rough reference.

Analysis has 27 titles (ratio 8.06) and computation methods 12 (4.62), against algebra's 2 (0.55) and wasan's 2 (0.52). These are probably not books for mathematics majors learning the field systematically, but a concentration of books for people heading into engineering or physics who needed to "be able to use calculus as a tool."

Just after the war, books with complicated diagrams were published by hand, on mimeograph and the like. (Most of them ended up in the Prange Collection.) Because they were rushed out, many of the books themselves say there is a lot that falls short.

In the present state of a defeated nation, there is much that cannot be done as one would wish, from the choice of type to the insertion of figures. We have been as careful as we could, but as we were in haste, some misprints are no doubt unavoidable. On this point we beg the reader's indulgence.

*Outline of Clinical Ophthalmology* (Rinshō gankagaku yōyō), Nakajima Minoru, Kurachi Yoshi, 1905-1982, Tohōdō, 1946

https://dl.ndl.go.jp/pid/1046601/1/4

Even so, it had to be put out — and one can take that as passion and good faith.

Physics (Class 42): 52 titles, ratio 3.23

Physics is the same: 15 in general, 11 in theoretical physics, 7 in electromagnetism — weighted toward the fields that bridge into electrical engineering (108 titles).

Chemistry (Class 43): 42 titles, ratio 3.44

In chemistry too, the hands-on side runs high: experimental chemistry (ratio 8.34) and inorganic chemistry (7.92).

Electrical Engineering (Class 54): 108 titles, ratio 2.26

Electrical engineering, at 108 titles, is near the top of every field; restoring the power infrastructure was one of the areas that needed hands the most immediately.

Language Study

The skew in language study can be read in this context. English, with 14 titles (ratio 0.85), is remarkably low. It was the language of the occupying army, so one would expect demand for study guides, and this is a strange point. Perhaps what circulated instead were small booklets and pamphlets at the everyday-conversation level; or perhaps, since English had never been entirely cut off even during the war, there was less of a rebound surge.

French, on the other hand, with 10 titles (7.81), and Russian with 7 (8.36), are abnormally high. This is probably less practical language study than a hunger for knowledge — a wish to read the enemy-language literature that had been forbidden during the war — or a reflection of intellectual interest in the Soviet socialist system.

Study Guides, Summed Up

Taken as a whole, what the study-guide data says is this: in an age when the institutions were not working, individuals went looking for the knowledge to survive on their own. School does not teach me, so I learn by myself; the state shows no direction, so I acquire a skill by myself.

Rather than schoolbooks for classroom education, this was probably a market of practical manuals for demobilized ex-soldiers and for young people whose schooling had been broken off by the war, who wanted to become engineers by self-study.

At the same time, hopes for students aiming at higher schools, and for children, ran high.

This book gathers together, in the order of the months, some of the observations and experiments you can do with the plants in your garden or out in the fields, month by month from the end of winter to the beginning of winter. Everything is explained on the basis of observations and experiments I carried out myself, one by one; and for the methods of observation and experiment especially, I have arranged things so that all of you can do them, and so that anyone can do them — using not only no microscope, of course, but as far as possible no magnifying glass either.

If you study science this way, I think that even people who dislike science will surely come to like it. If every one of you, without exception, comes to like science, then the Japanese from now on will all have scientific heads. Only then will Japan become a truly civilized country and be able to join the ranks of the civilized countries of the world. What a joyful thing that would be.

*Plants Through Twelve Months: Observations and Experiments* (Shokubutsu jūnikagetsu: kwansatsu, jikken), Harada Mitsuo, Kodomo no Kuni Foundation, 1946

National Diet Library Digital Collections ↗

Unfortunately, the Japanese did not all end up with scientific heads.

Notes on the Data

There are differences between the two datasets used in this analysis. The main ones are below.

Different NDC editions. Russian and Soviet literature (98), Italian literature (97), and literature in other languages (99) do not exist in the overall data. On the overall side, literature (Class 9) totals 214,824 at one digit against 209,853 at two digits, and the roughly 5,000-title gap is presumed to correspond to these classes. The cause is likely a reshuffling of the classification scheme accompanying an NDC revision (consolidation into 96 and below, etc.), or a difference in how faceted search is specified. The ratios for classes 97 through 99 therefore have nothing to compare against and are invalid.

Inconsistent notation of names. The just-after-the-war data uses the middle dot "・" as a separator (e.g. "Manners and Customs・Folklore・Ethnology"), while the overall data uses the Japanese comma "、" (e.g. "Manners and Customs、Folklore、Ethnology"), which indicates that the data was retrieved at points when the NDL API or facet versions differed.

The Prange Collection being excluded is truly a shame.

Problems with This Analysis

1. Asymmetry of the comparison

I am comparing "one and a half years" in 1945-46 against "the whole period (1.58 million titles)." That "whole period" includes fields that exploded after the postwar high-growth era (new technologies such as information engineering, contemporary fiction, subculture-related material, and so on). So what is being measured may not be "the needs of the time" but "the difference from the publishing structure of the present day."

2. "Demand" versus "restricted supply"

Paper was short at the time, and publishing was under a paper allocation system run by the government and GHQ (CIE: the Civil Information and Education Section). "Enlightenment in scientific thinking" and "literature of democratic nations (in translation)" fit GHQ's occupation policy and were given priority in paper, while kokutai ideology and the martial arts got none. The skew in the ratios may be the result of censorship of publishing by those in power.

3. Survivorship biasBecause the Prange Collection is excluded, the 3,564 titles extracted are weighted toward "books that went through the proper legal-deposit route" and "books the NDL was able to collect in later years." It is also highly likely that much of the guerrilla publishing — kasutori magazines, shoddy pamphlets, underground publications — has dropped out of the data entirely.

Conclusion

Right now I am trying to read every preface and afterword in the 6,037 search results for publications issued between 1945/06/01 and 1946/09/30 that are held in the National Diet Library Digital Collections. I have come to some sense of what the year after the defeat was probably like, but I wanted to use objective data and check whether that sense was right. So I decided to analyze the skew in the composition of the NDC (the Nippon Decimal Classification) classes. My "isn't it something like this?" is, I think, broadly correct — that is the conclusion as things stand.