Facts: An Obituary

The other day, RTÉ broadcast a short obituary of Martin McGuinness, following his death.

It contained the following paragraphs, which is also published on RTÉ’s website.

(McGuinness) quickly rose through the ranks of the Provisional IRA, as its campaign of violence targeted unionists, and Catholics who did not toe the line.
Bloody Sunday, which resulted in the deaths of 14 men and boys, bolstered support for the IRA and McGuinness.

And then, further along, narrating events from the 1990s:

As the countless deaths continued, the IRA began to realise that it was not going to end this by military means. A ceasefire was announced and negotiations began, resulting in the Good Friday Agreement.

I want to test the claims made here, and explore what the effects are of making these claims.

First, we should acknowledge that it’s difficult to write a short narrative that encapsulates a long sequence of turbulent events. It’s inevitable in these circumstances that some important detail will be lost. What’s more, opinions will differ, depending on political perspective, regarding what is important. It may be all the more likely, given this, that what is intended as an impartial account ends up revealing the -inevitably- partial perspective concealed in this apparent impartiality.

In testing these claims I am not proposing that there is an Olympian, ‘objective’, ‘bias-free’, viewpoint from which such events can be reliably related. Rather, I will illustrate, as I go on, that such a viewpoint is ultimately impossible. I do not contend that there is no such thing as facts. Things do happen. People die, and other people kill them. But how those facts are interpreted, and acted upon, create other facts.

The first claim I will address here is that the deaths in the conflict were ‘countless’. These deaths were counted.

They were counted, for example, in David McKittrick’s book Lost Lives, and in Malcolm Sutton’s Index of Deaths, available on the CAIN website, and also published in a book. These counts are not and could not be altogether accurate, since the full circumstances behind many deaths are not known, and so the classification in some cases -for example, when infiltrators are involved- may not be clear cut. RTÉ’s claim may well be a loose turn of phrase, rather than a factual statement, but such a loose turn of phrase disregards the work people have done in counting them, and in recording the circumstances. Not everyone is so indifferent.

The next claim is that the IRA’s ‘campaign of violence targeted unionists, and Catholics who did not toe the line’. To address this I will use the Sutton data on the CAIN website.

Processing the data

The presentation of the data on the website does not appear to have changed for a long time. It looks as though it was published in the late 90s or early 00s. I was unable to find a single data file that could be used for a thorough and intelligible analysis. So I made one.

The date range I chose was from the start of the index to 31st August 1994. I may include the entire index in the future; it was just that here I was concerned with the sequence of violent events narrated in the obituary, which culminates with the IRA ceasefire.

The detailed data on the Sutton Index is divided into a page for each year. It all follows the same structure, detailed here.

I pulled all the entries into a single text file, then did some parsing in Python: regular expressions to identify the variables contained in each line, then appended the variables from each line to new lists, in terms of the date, name, description, the person’s status in the conflict (e.g. civilian, UDA, IRA etc), the group responsible for the death, and the religion of the dead person. I then combined the lists and exported to a csv file, and did the visualisations using Tableau.

I had some unexpected reflections when working with the data. I do this kind of thing fairly regularly, but it doesn’t normally entail records of violent death. I had a fairly clear picture in my head of the kind of data I wanted to present to begin with. But I found myself wondering whether I wanted to do the visualisation at all. You see, so much of data visualisation technique, with the range of tools now available, emphasises making things both clear and aesthetically pleasing. Sometimes, people give so much emphasis to the aesthetics that the information they are trying to convey takes on a lesser importance. It becomes more about making a show of being able to present the information than conveying the importance of what is being presented. When you’re making graphs about people who have died violent deaths, that kind of trade-off feels macabre.

When I was parsing the text file, I noticed that some of the lists I had put together were not the same length. There were some lines going into the wrong list. I started to debug the code. The reason the lists were of different length was because there were some lines that related to children. People’s ages are contained in brackets in the source file and I had written a regular expression, using the location of the brackets, to classify the line in terms of whether it was their name or a description of how they died. But I got the code for the digits wrong, and though I was classifying adults and teenagers correctly -their age had two digits- younger children had only one. So when I was testing to see if the code would work, and it repeatedly would not, I kept seeing the same name appearing in the death description list:

Ball, Jonathan (3)

He was killed, along with Tim Parry, by IRA bombs in Warrington on 28th of March 1993.

Each record has an entry for religion. Each person from Northern Ireland is classified according to a religion. Others not from Northern Ireland are not. I also felt unease about this. A friend of mine is on the list, classified as Catholic. The classification is accurate, but I wondered about the importance attached to this, the way it formed part of the record in every case, as a relevant datum.

Why religion, and not hair colour or favourite TV programme? We know the answer: in many cases, people were singled out and killed based on what their religion was. What is more, when you see that a person was of a particular religion, you can better gauge the likely effect of that on the wider community.

However, the label is uniformly applied to everyone, regardless of how much religion was a part of their life, and regardless of what kind of relations they might have had with other people of other religions. I understand the rationale behind the classification, and in fact I use the classification further down. Nonetheless it feels like a sectarian logic imposing itself on the data. Just because we can make things clear cut does not mean that they are clear cut.

Nor is the status variable without its problems. The broad groupings here are civilians, British Security Forces, Republican Paramilitary Groups, Loyalist Paramilitary Groups, and Irish Security Forces. Some kind of classifying logic is inevitable if you want to understand who killed whom and with what motive, and yet the application of the logic seems to bring with it certain assumptions and exclusions, a tendency or at least a temptation to treat certain deaths as better, certain lives more expendable, than others.

The point here is that what we offer up as ‘facts’ are the product of subjective judgment. This is worth bearing in mind the next time someone claims we are living in a ‘post-fact’ era, as if there were a previous era in which the ‘facts’ were more solid. At best, they had the appearance of greater solidity.

‘Targeting unionists’

Did the IRA’s campaign of violence ‘target unionists’, as RTÉ claims? It depends on what you mean by ‘unionist’. If, as per the graph above, British Security Forces -the category used by Sutton- are unionist by definition, and there are plenty who would argue that they are, then RTÉ’s claim is correct. However, it is hard to imagine an RTÉ news report, during the years of the conflict, that said the IRA had shot a unionist soldier dead. This is because ‘unionists’ means something else, generally speaking. It would be far more accurate, from this data, to say that the IRA targeted British Security Forces.

But we need to be careful about the use of the word ‘target’ in relation to this graph. Consider the third category, ‘Republican Paramilitary Groups’.

Here, the Sutton data shows that the IRA was responsible for the deaths of 132 of its own rank. When we inspect the data, however, we find in the description that 100 died in premature bomb explosions. We can hardly say that they were the target of such explosions.

This raises the question of the extent to which the IRA targeted civilians, and unionist civilians in particular. This is how the data looks when it is broken out by Sutton’s religion category.

But to ascertain whether the IRA targeted unionist civilians, we have to accept to some degree that to be Protestant means you are likely a unionist, and that to be Catholic means you likely are not. Once more, we end up in murky waters. How appropriate is the word ‘target’?

Sutton’s descriptions contain the phrase ‘Inadequate warning’ when someone was killed in a bomb where inadequate warning was given. Or ‘Mistaken’ when someone else -usually a member of the security forces- If we include this, we get the following:

So there are certain instances where ‘target’ is hardly the appropriate term. To point this out is not to provide justification but to stress the lack of precision in terminology. And this scarcely matters to the victims. This is not to say that Protestant -or unionist- civilians were not targeted in attacks. Many were, and others were killed in attacks that targeted family members. The short descriptions provided by Sutton in this regard are gruesome.
The overall picture, however, is not one of a force that ‘targeted unionists’, but one that primarily targeted British security forces. Many unionists, of course, are likely to see things differently.

What about Catholics who did not ‘toe the line’? Sutton’s descriptions contain the phrase ‘Alleged informer’, where such information was available. The truth of the claim depends on what you mean by ‘Catholic’ and what you mean by ‘not toeing the line’. If, for instance, a British asset within the IRA claimed that a certain person was an informer, in order to maintain his own cover, does that mean such a person did not ‘toe the line’? There are lots of grey areas here. The broader historical record, however, does not suggest that the IRA targeted Catholics because they did not support the IRA. If this were the case, a lot more would have been killed. It is worth recalling that at the time of the IRA ceasefire, most Catholics did not vote for Sinn Féin.

The next claim is contained in RTÉ’s narration of events that led to the IRA ceasefire. This needs to be put into context. There is no account in the obituary of any other force being involved in killing. Not the British Army, not the RUC, not loyalist paramilitaries. A person who did not know anything about the conflict would be led to believe that it was only the IRA that was the belligerent. However, the data tell a different story.

The graph above shows deaths by year, colour-coded in terms of the group responsible. What is striking -to someone born then but with no memory of it- is the intensity of the carnage from 1972 through 1976, relative to later on. It is not as if later on was a walk in the park. What is also clear here is how RTÉ’s presentation omits the role of the British Security Forces and Loyalist Paramilitaries (to use Sutton’s categories) altogether. These groupings had no ‘campaign of violence’, by RTÉ’s lights, despite the fact that the British Security Forces had provided arms to loyalists, had conducted internment that initially left loyalists free, and had maintained the UDA as a legal organisation until 1992.

If we examine only civilian deaths alone, as shown above, the picture is even more stark. It is Loyalist Paramilitaries that are the main grouping targeting civilians -and, as we know, through work done by people like Anne Cadwallader, doing so with the aid of the British State. The data for the entire period is summarised below.

Particularly worth noting is the sharp rise in the civilian death toll from Loyalist Paramilitaries from 1990 onwards, the period marked in RTÉ’s narrative with ‘as the countless deaths continued’.

By the time of the IRA ceasefire, the civilian death toll was escalating year-on-year, and Loyalist Paramilitaries were killing nearly three quarters of civilians. Loyalists at the time, such as Billy Wright, were quite open about their modus operandi: as long as the IRA was at war with -as they saw it- the unionist community, they would serve up Catholic civilians in body bags. What is not yet clear, however, is whether this modus operandi was seen as beneficial from the standpoint of the British State. But these considerations have never mattered to RTÉ, or to Ireland’s media and political establishment more broadly. There are always some facts to be created, and others to be given proper burial.

UPDATE 26/03: I have written a follow-up post, considering some of the responses to this post, and some further aspects of the data and how it is interpreted, here.