The R word

Reprisals: all mentions, 1939-1945

The word 'reprisals' popped up during my 1940 post-blogging quite frequently. After one post I had the idea of checking whether it could be used as an index of British attitudes towards the bombing of Germany throughout the rest of the war. The short answer is: not really. But it was still worth trying.

With The Times and the Manchester Guardian/Observer databases I can luckily do this in a semi-automated fashion. Automated because I can do keyword searches on the full text of the newspapers, semi because the interfaces are crude and require manually stepping through the date range to bin the data. For example, searching for the word 'reprisals' in The Times database between 1 and 31 July 1940 gives 16 articles; doing the same between 1 and 31 August 1940 gives 18 articles; and so on. I then put these numbers together and plot the results.

So above is the plot for the whole war, binned into quarters (January-March, April-June, July-September, October-December). The numbers represent at least one use of the word 'reprisal' in an article. This is a problem: how articles are defined varies a bit depending on which database it is and what part of the newspaper is being searched; several letters to the editor on diverse topics can be bundled into one 'article', for example). Blue is for The Times, and red the Manchester Guardian/Observer. That's the second problem: the Guardian and the Observer are lumped in together in the ProQuest database I'm using, simply because since 1993 they both have had the same owner. Not only did the two newspapers not have anything to do with each other in the 1940s but they also had very differing political views, the Guardian being liberal and the Observer conservative. Having them in the same dataset like this is thus quite problematic. Because the Observer was published only on Sundays I could have filtered it out at the price of making the whole process more labour intensive, but it's probably not worth the effort anyway.

That's because of the third problem, which is really the problem: reprisals ain't reprisals. Without knowing the context it's impossible to say whether a mention of 'reprisals' in September 1940 meant the same thing as a mention in July 1944. In fact, they almost certainly wouldn't have. It says nothing about whether the comments were 'pro' or 'anti' reprisals. The word could be used in non-military contexts too, of course. So the above graph says nothing particularly interesting in itself. It confirms that the most chatter about 'reprisals' took place during the latter half of 1940, that is to say during the Battle of Britain and the early Blitz, and also that with a few exceptions the chatter seems highly correlated across the press. But it is only illustrative and suggestive: it doesn't obviate the need to actually go and look at the primary sources to see what was actually going on.

That said, what do the peaks in the graph represent? The first one is in late 1939, and refers to economic reprisals taken by both sides: the impounding of enemy contraband in neutral shipping, for example. After all the reprisals talk in 1940, it became a relatively neglected topic until 1942. The contexts there are all over the place: the German press described the Baedeker raids as reprisals for the bombing of Rostock and Lübeck; it came up in relation to collaboration and resistance in occupied Europe (including the assassination of Heydrich); Dr I. Schwarzbart said that only 'immediate reprisals' could deter Hitler from completely annihilating Poland's Jews, a million of whom were already dead (The Times, 30 June 1942, 2). In October, German authorities put British prisoners of war in chains, an act forbidden under the Geneva Convention; both this and Britain's mirror response were described as reprisals. The final peak, in the third quarter of 1944, had nothing to do with V-weapons; instead it was due to the anti-partisan actions taken by the German army after D-Day.

Reprisals: all mentions, July 1940-May 1941

Let's turn then to the Battle and the Blitz, this time looking at the numbers for each month. This shows that reprisals were widely discussed in The Times and the Guardian/Observer only during the first two or three months of the Blitz. Interest fell off during the winter months, declining to practically nothing in February 1941 (why this would be is an interesting question in itself). It was starting to increase again by the time the Blitz ended in May.

Reprisals: letters to the editor, July 1940-May 1941

While there's a lot to complain about with regards to the interfaces these databases have been saddled with, one good thing is the way you can narrow down searches to particular sections of the paper, e.g. the leading articles or letters to the editor. Doing this filters out the 'news' and gives us 'opinion': what news editors and readers felt was important enough to give their opinions on. The above plot shows this for the letters columns. I'd already noted during the post-blogging that the Manchester Guardian's letters columns were almost silent on the reprisals debate. It looks like this changed in October; and from then on its readers were generally as interested in reprisals as those of The Times, if not more. (In this case I can omit reference to the Observer, as it didn't publish correspondence from readers.) Again, this doesn't tell us why they suddenly became so interested (or, perhaps, why the Guardian started publishing their letters).

Reprisals: leading articles, July 1940-May 1941

This one is just the leading articles, and shows that none of the editorial writers were particularly interested in the reprisals question, usually with no more than one or two references per month for most of the period. The big exception here is the Guardian and/or Observer in September 1940, with a whopping seven articles. On closer inspection, two of these refer to Vichy reprisals for Dakar. Another four refer to the German claim that the raids on London were themselves reprisals for British attacks on Berlin. Only one (in the Guardian) in fact says anything for or against the idea of launching air reprisals against Germany. (Which explains why I didn't notice this 'peak' when I was post-blogging!) It's this sort of uncertainty which makes trawling these databases in this way a fairly fruitless exercise. If they had more powerful search functions (or, in the case of Cengage's Times archive, one that actually worked properly), it might be possible to frame more sophisticated queries to winnow out more of the chaff. But for now the traditional way of doing history ('reading the primary sources') still works best, at least for me and in this area.

CC BY-NC-ND 4.0 This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License. Permissions beyond the scope of this license may be available at

5 thoughts on “The R word

  1. Derek R

    Interesting way to track the gestalt, by a simple keyword search. It's sort of like googles zeitgeist. Makes you wonder if the lingo changed from "reprisals" to a different word, like "revenge" for example? Too bad the interface is clunky, you could pull out the thesaurus and track all sorts of "reprisal-equivalent" words.

  2. Post author

    You're right, it's definitely possible that a different word might have been used at different times. Especially since the way the bombing of Germany was framed would have changed -- in 1943, for example, it wouldn't have been justified as reprisals but as military necessity, preparing for invasion, etc. So again you really need to know the sources first to know what words are significant and why.

    Alternatively, if one had access to the database itself you could be more creative and look at words which appeared frequently in relation, say, to 'berlin' and 'bombed', and other sorts of clever relational things. And then if more newspapers were available and in a consistent format you could repeat that across all of them. That might give interesting results.

    Of course there's the Google News Archive, but it has its own problems, for example lumping in every newspaper everywhere into the search (you can narrow it down to one specific newspaper, if you know which ones are in there). And Google doesn't seem to have released an API for it -- otherwise an enterprising person could write a better interface. But just for the hell of it, here's their take on the frequency of 'reprisals' between January 1940 and December 1945 -- it even does the graph for you, and you can zoom in to find more specific periods of interest.

  3. Reading a Google News Feed for World War II was a surreal experience.

    And all that reprising was depressing. Where are the slideshows and celebrity news? (Top 10 Places to Jitterbug the Night Away, "Bogart and Methot Splitsville?)"

  4. Interesting. I've tried something similar with occurences of different cognates for Germans and Nazis across both world wars, and I wonder if this works better for terms which are less multi-usable than 'reprisal'? 'Fritz' and 'Jerry' both cause problems, of course, because they throw up everyone with those proper names, but I think you can plot an increase in the use of 'Hun' over the course of both wars (and a decrease in the occurence of 'Nazi' during WW2) which seems less ambiguous (although, given newsprint rationing, shorter words were probably more popular anyway!)

  5. Post author


    It would probably be possible to do a fully automated aggregation post-blog, pulling (say) all the front pages out of Google News Archive each day for X number of years ago. Someone clever should do that.


    I think you're right that this approach would work better for words with more specific meanings. Conversely having a very narrow frame of interest, as here, doesn't help. If I'd just been interested in whether talk of 'reprisals' generally increased in wartime in whatever context (as you'd probably expect) then that would have worked.

Leave a Reply

Your email address will not be published. Required fields are marked *