A Textual Analysis – The Rise and Fall of the Third Reich

William Shirer was a foreign correspondent with UPI and CBS stationed in Europe before and during World War II. After the war he wrote The Rise and Fall of the Third Reich which was first published in 1960. As the title suggests, the book covers Hitler’s and the Nazi’s rise to  power through their downfall at the end of World War II. It is a brutal and agonizing journey.

Here is Shirer’s closing:

The guns in Europe ceased firing and the bombs ceased dropping at midnight on May 8-9, 1945, and a strange but welcome silence settled over the Continent for the first time since September 1, 1939. In the intervening five years, eight months and seven days millions of men and women had been slaughtered on a hundred battlefields and in a thousand bombed towns, and millions more done to death in the Nazi gas chambers or on the edge of the S.S. Einsatzgruppen pits in Russia and Poland – as the result of Adolf Hitler’s lust for German conquest. A greater part of most of Europe’s ancient cities lay in ruins, and from their rubble, as the weather warmed, there was the stench of the countless unburied dead.
No more would the streets of Germany echo to the jack boot of the goosestepping storm troopers or the lusty yells of the brown-shirted masses or the shouts of the Fuehrer blaring from the loudspeakers.
After twelve years, four months and eight days, an Age of Darkness to all but a multitude of Germans and now ending in a bleak night for them too, the Thousand-Year Reich had come to an end. It had raised, as we have seen, this great nation and this resourceful but so easily misled people to heights of power and conquest they had never before experienced and now it had dissolved with a suddenness and a completeness that had few, if any, parallels in history.

A painful part of our collective history.

To state the obvious, this book is filled with a multitude of characters and I thought it would be interesting to see who plays large rolls and, since the book (for the most part) proceeds chronologically, when the characters come on and off the stage. This led me to parse and analyze the text. I parsed the 1990 edition which is 1,029 pages. Below is a textual analysis of this book.

I parsed and analyzed the text using python and pandas. I used a superset of stop words from here and here. The graphs are in plotly. And the word clouds were made using Andreas Mueller’s generator. All of the code is in github.

Some basics about the text:

• Number of words: 571,387
• Number of words (sans stop words): 244,881
• Number of unique words: 22,748
• Number of unique words (sans stop words): 22,266

Here are the top 15 non-stop words and their number of occurrences.

Sounds about right.

And here is a word cloud for the book as a whole.

The book is split into 31 chapters with an Epilogue and Afterword.

For the most part the book follows a chronological order, although not entirely.

Here is a view of where occurrences of specific years occur in the book. This gives a sense of where chapters fit in the chronology of the era.

https://plot.ly/~yg2bsm/106

I made a list of the key people and places from the book and then cross-referenced them against the text to see where they occurred. I made these 2 lists manually (for the most part). I tried to use Spacy to identify the “entities”, but it struggled with the multitude of unique names (for both people and places).

Here are the top people and places that occur most frequently in the book.

Note that the count for Hitler includes “Fuehrer” (its 4237 without) and the count for Mussolini includes “Duce” (its 333 without).

Here is the word cloud for the occurrences of people in the book.

And here it is for places. In this instance I treated the cities and countries separately.

I thought it would be interesting to see which of the different players had the biggest roles across the different chapters. Below is a table showing the individuals that appeared most often in each chapter (most often is to the left and the fifth most often is to the right).

You can see that Hitler (+Fuehrer) is the #1 person in each chapter except for 1. But you do see quite a few different names. 56 different people are in the top 5 of at least one chapter.

The table below shows the top individuals that occur in the top 5 of the most chapters.

No surprise about Hitler. Ribbentrop was the Nazi Foreign Minister from early 1938 through the end of the war. Mussolini was the Italian fascist dictator / Hitler ally. Goering, Goebbels, and Himmler were all early Nazis and served as head of the Luftwaffe (German Air Force), Minister of Propaganda, and head of the SS respectively. General Halder was Chief of the German General Staff. Paul von Hindenburg was President of Germany from 1925 until his death in 1934 (whereby Hitler decreed himself leader). Franz von Papen was appointed Chancellor by Hindenburg in 1932 and then became Vice-Chancellor when Hitler took over as Chancellor in 1933. He resigned in 1934. Vyacheslav Molotov was the Soviet Minister of Foreign Affairs from 1939 until after the war.

Back to the book as a whole. Here is a graphical depiction showing where the most frequent words (people / places) occur over the duration of the book. My WordPress theme is not giving the charts much in the way of width, so there is a link to a full screen version for each one.

Below is the people graph. Click here to see the graph full screen.

The greyed-out names are hidden, but can be turned on by clicking the name. Hitler is turned off by default because if it is shown then the scale is such that the curves for the other names appear near 0.

You can see the Chamberlain spike around the 40% mark. This corresponds to the time surrounding the Munich Conference in September 1938 which culminated in the infamous Munich Agreement which led to the Nazi occupation of Czechoslovakia.

From the Munich Conference: (left to right) Mussolini, Hitler, Schmidt (translator), and Chamberlain

You can also see the Mussolini spike towards the last 10th of the book which corresponds to his downfall (pushed out of power and arrested by the Italians).

Below is the places graph. Click here to see the graph full screen.

The occurrence count for the place names include the cities within in their counts. For example, the count for France includes the count for Paris. Again, the greyed-out names are hidden, but can be turned on by clicking the name. Germany is turned off by default because if it is shown then the scale is such that the curves for the other names appear near 0.

The largest spike is for Poland which corresponds to the Polish Invasion and the onset of World War II.

Similar graphs showing the occurrence by chapter (instead of by a linear percentage-based splittage) were created and can be seen here: People, Places.

And finally, here is a matrix of word clouds (one cloud per chapter) for people followed by places. This is another interesting way to view the evolution of the major players and locales over the course of the book (and time period).