What types of conflict are taken into account in the Uppsala Conflict Data Program and how do you define armed conflicts?

We try to distinguish three different types of organized violence, which we think are often mixed together, causing a lack of necessary clarity. The first basic type is armed conflict, which is defined as a political disagreement between one actor (a state) and another actor (which could be a state or an organization). That conflict must have arrived at a certain magnitude, and that we measure in terms of 25 people being killed during a year. It means we monitor all kinds of situations and see how they develop; when they reach the threshold of 25 deaths, we include them. That's our basic concept of armed conflict. The key word here is actually the disagreement, or as we call it in slightly more academic terms, ‘incompatibility’ – there must really be a political disagreement here. We would not include, for instance, skirmishes, guards shooting at the border and situations like that, when there seems to be more of an accident, or where it is definitely not tied to a political strategy.

What if you had, for instance, shootings between the armed forces at the border between states – would that be considered, in your concept, an armed conflict?

Perhaps – we would look at it to make sure that it really comes from a political design, and that the guard did not just happen to be shooting. And, of course, there have to be more than 25 people killed. This threshold means a number of smaller incidents are not included on our annual list of armed conflicts.

Apart from this basic concept of armed conflict, what are the two other types of conflict that you examine?

There are many conflicts going on between non-state actors: communal violence, tribes attacking each other, gangs of various sorts, etc. We try to capture these, but as a separate category, which we call non-state conflict. It is often unclear what motivates this violence, or how political it is – therefore, we see this as a different category from armed conflict, which is clearly a politically directed fight about controlling the government or the territory.

This type of violence is not against the state, but really inter-communal violence – would the case of Somalia fall under this category?

For quite a long period of time, yes. It varies depending on when there is a government, and when there is not. But a lot of the violence in Somalia would be of this non-state type, warlords fighting each other.

And the third type of conflict?

The third category, we call one-sided violence – that is, when violence is aimed at particular populations which are not organized. This would cover, for example, genocide, many acts of terrorism and so on. This violence can be perpetrated by a state or by a non-state actor.

We find that these three definitions are fairly distinct from each other. This makes it possible to analyse whether the three categories combine or influence each other. This is how we try to capture a number of the issues, which are not regular armed conflict, in a significant way.

When dealing with terrorism, you have the acts of terrorism, which would be classified as ‘one-sided violence’, as you defined it. On the other hand, there are counter-terrorism mechanisms. Would counter-terrorism activities fall under the first category of armed conflict, or would they also fall under the third category of one-sided violence?

Most counter-terrorism would probably be under the first category, because normally – for instance, with Al Qaeda – a government is fighting an organized non-state actor. So we would define the United States fighting Al Qaeda, wherever the fighting is located, as an armed conflict.

The starting point of distinction between your different categories is the concept of armed conflict. Would you say that this is very close to interstate conflict?

The history is important. I was studying a project at the University of Michigan called ‘Correlates of War’, which focused on interstate conflict. At first, it defined different types of interstate conflict, but then later created a separate definition for internal conflict. I felt that there should be one definition of armed conflict, so we at Uppsala tried to integrate them so that they became comparable. You can do this with our current definition. If you have an armed conflict, involving one state against another state, that's an interstate conflict. If you have a conflict involving one state or government versus a rebel movement, that's an internal conflict. In that way, we have the same definition for an armed conflict. It's based on the same criteria whether it is interstate or internal.

The question of occupation is obviously a political disagreement or incompatibility – would that fall under your definition of armed conflict?

Well, that would depend on who the fighting actors are. For instance, Israel versus PLO – that's not an interstate conflict because Palestine is recognized by some but still not an independent state. It is classified as a state versus a non-state actor, which definitely falls under our first category of armed conflict.

This first category would be very close to what is actually foreseen in the international law of armed conflict, where there are basically two types of conflict – international and non-international

Yes, that's true. The idea is that you should not have very strongly differentiated criteria for what is an armed conflict. Today, when things are much more mixed, it's helpful to have this definition, which looks at which parties are fighting. The problem is, of course, the international law system is still very interstate-oriented.

In the first category, a political disagreement is required. However, there is now often a mingling between political factors and criminal elements which are linked to the conflict – for example, drug-related criminality – which fuels the conflict, like in Colombia or in some African countries. Would such a situation, where there is an intertwining of political and criminal motives, still be considered an armed conflict?

There it becomes really very tricky. For us, the key thing is the disagreement, the incompatibility. Do these criminal groups aspire to take control over the government, or not? Do they aspire to take control over a particular piece of territory? If they do, then we would include them. Some drug-related conflicts, in Colombia for instance, are conducted to influence the political process. But a lot of it is not politically motivated – sometimes actors are not fighting there for political power, but for criminal gain of various sorts. We would not call this an armed conflict, as the actors do not want to exert political power. We would separate political conflict from criminal activities – we don't want to have pure criminal activity in the category of armed conflict.

Let us take an example of drug cartels who are violently opposing military forces, but who clearly do not have political aspirations, at least at this stage. Would you consider this to be in the first category?

Generally, no. When for instance looking at Mexico, we consider it more as non-state violence: two gangs fighting each other for control over a particular trade in a particular city, for example. I don't think we have seen any of it, as yet, really politically challenging the state. Drug lords are not trying to take power, they do not want to be running the state. Maybe they want to impact on the state, in the sense of making sure it does not interfere with their business, but it's not the same as trying to take political control. We try to distinguish this violence according to what kind of aspirations these groups have.

So it doesn't matter if groups are fighting for, let's say, religious or economic reasons – the determining factor would be their actual or desired impact on the political system

Yes, that would be the key issue. We are not saying anything about the causes of the conflict. We are basically just presenting organized violence which can be caused by many different factors. Sometimes it can have an ethnic background, sometimes a class background, etc. The cause is separate from the actual phenomenon of the fighting.

This still seems a rather traditional international law approach: a distinction between the laws regulating the reasons for going to war (jus ad bellum) and the situation of war afterwards (jus in bello)

That's right, and I think there is clarity to this – it is much more intellectually satisfying to separate the causes. Many people say, for example, that conflicts are all ethnic conflicts, but when you start to look at them, a conflict is rarely related to only one factor. They are always a mixture of issues, but there is still a conflict going on that we can record. To understand its origins, functions and solutions is an analytical issue that is separate from recording the phenomenon.

You mentioned that you include both interstate conflict and conflicts between a state and a non-state actor, which must involve a political disagreement. Do you make any distinction between international, non-international and internationalized armed conflicts in this category, or is it the criterion of political disagreement which interests you primarily?

The information we have available can be used according to the analysis you want to do. For instance, we can also record which other actors are involved in a conflict. There would be the primary parties (the state and the opposing state or organization actually fighting). Of course either party can have support from other actors, which we call secondary parties. You could have a secondary warring party sending troops to the area – if this party is another state, then the conflict is internationalized. There are a quite a number of such conflicts. We have done a separate study on this. The secondary parties may not necessarily be involved in a war using their own troops, but supporting one or the other actor financially, politically, or in other ways. You could have a whole set of secondary actors involved, and we record that.

In our database, we have about 120 different dimensions available – for free – for anybody to do his or her own analysis. For instance, in the Journal of Peace Research we regularly publish categories such as interstate, internationalized, and purely internal conflicts. We want this to be easily accessible, for whatever categorization you want to make as an analyst. However, we are not saying that a particular conflict is only of one particular nature. To me, every conflict is very multi-faceted. What we can do – intellectually honestly – is not to say: ‘this is an internationalized conflict’, but rather to give the information and let the analysts make up their own minds.

In the second category – non-state violence – what are the major types of situations you look at?

Well, these would be situations where the state is not strong or not directly involved. The typical example, as you mentioned, would be Somalia. Another example would be Nigeria, where one confrontation is normally described as taking place between Christian and Moslem groups; or in India, what is known as communal violence. These events can be very devastating and affect people very strongly. But our record shows that they often do not continue for long. There are outbursts of violence that will last for a couple of days, or maybe weeks, but they are not protracted and do not go on like a war. They are often stopped or contained. They have different dynamics, but they definitely should be part of our picture of political violence.

Could this also include violence in favelas, in Rio for example?

That's the kind of issue we have been debating – exactly what to do there – but in principle that should be included, if we can determine that the groups are clearly organized in a comparable way. That is a non-state conflict. Or for example, if you have armed gangs moving in towards the native population, that may be more like one-sided violence.

In the third category – one-sided violence – the most striking example is certainly the case of genocide, but you mention that acts of terrorism could also qualify. Would these be Al Qaeda-type situations?

Yes, the idea here is that it is not necessarily a political battle between organized actors on each side. One side is organized and deliberately targeting civilians. The typical example is, of course, the attack on the World Trade Centre in New York in September 2001. The victims were civilians with no military function, and who were not part of the political administration of the United States. They were doing their regular business and were suddenly exposed to this attack and killed. Terrorism is really very much of this nature – an attack on civilians who are not necessarily involved in a political battle. The same with genocides – for instance, in Rwanda, in 1993–1994: there was a small armed conflict going on, but then there was a huge genocide, which we view as a separate issue. The genocide, of course, killed maybe 800,000 people, whereas the armed conflict was very limited. The odd thing, however, was that the United Nations, for instance, basically focused on the armed conflict rather than the genocide.

At one stage at least, before the attacks of 9/11, Al Qaeda was an organized entity – at least in Afghanistan itself, where they had clearly organized structures. Now, it's probably a rather loose network, with merely individuals acting. Would that then move from the first category to the third category?

Yes, we look at each situation, and even divide it up. If you have battles between the armed forces and an organization like Al Qaeda in Mesopotamia, this would probably be an armed conflict. If the same organization is doing suicide bombings on civilians, that would be the terrorist part of it.

It's particularly important concerning the jurisprudence in the United States, where the Supreme Court clearly said in Hamdan decision that there was an armed conflict going on between the United States and Al Qaeda

We have defined the US and Al Qaeda as the opposites in one-armed conflict, because that fits with our armed conflict definition.

Battles do take place between the two organizations in Afghanistan, but if the same happened in Yemen, for example, could you talk about an armed conflict there?

Yes, because the key thing is that they are in a disagreement. There is an incompatibility, so it doesn't really matter where the battle in itself takes place; it's the same organizations fighting on different battlegrounds, so to speak.

But even so, the individual actors in such situations may have a very loose, maybe only philosophical, link to Al Qaeda, and may not actually be organized under a central command

Now we are into the key thing. That's what we would like to try to determine – is this really organized by the centre, Bin Laden, wherever he is? Or is Al Qaeda just a sort of inspiration for these other groups? In principle, we look at who is acting, and whether they are really a part of Al Qaeda or a separate body. Sometimes, like Al Qaeda in Mesopotamia, we will define it as a separate new body, rather than part of any strongly centralized structure.

So it depends on the situation – for example, the classification may be different in Afghanistan, where the group is more organized, as opposed to the Madrid attacks, where the attackers were acting more individually?

Probably these groups are separate, but inspired by Al Qaeda. You find this with many actors. There is a real problem, for instance, with Palestinian groups. New ones appear now and then, and it is very difficult to know if they are part of Hamas, part of Al Fatah, or on their own. That's always a difficulty in some of these situations where opposition groups are fairly fragmented. It's also hard to determine how centralized some of these organizations are. The idea for us is to try to include as much of this organized armed violence as possible. We have no reason to exclude it for any other concern except wanting our definitions to be applied strictly.

In the course of your studies on conflict, did you note new types of conflict emerging, or that some types were becoming more prominent?

That was our idea in the beginning. Many people were saying that we no longer have this old kind of armed conflict, and that now we have new wars which are more of the non-state or terrorism type. What we are trying to do in the project right now is to gather data on the latter for the past 20 years to see if there is such a difference. From some of these results, we have found that armed conflicts are special, because they are politically driven by the state, have a longer duration, and are mostly better financed. They are more continual and consistent than non-state violence – which tends to flare up and disappear much more rapidly – or one-sided violence which is about targeting and trying to get spectacular effects, making people fearful but without continuously fighting in the same way as in an armed conflict. It seems these three categories capture different kinds of conflict phenomena. Still, the finding is that armed conflicts are really those drawing in the most resources, resulting in more deaths and having the largest effects on the security of people.

Is this mainly due to the protracted nature of these conflicts?

Yes, and the resources involved. Take the war this year in Sri Lanka, which is the most devastating we have seen for several years. You can see that the parties are two strongly armed actors who have been fighting each other for more than 25 years, and the devastation seems to be tremendous. Even with terrorist deeds like the World Trade Centre, you don't come up to these enormous numbers. The fear effect of terrorism is very strong, but the actual killing is probably less than in armed conflict.

However, the budgets reserved for dealing with traditional threats and conventional warfare are often decreasing while the part of the military budget reserved for unconventional warfare steadily increases

Yes, and I think it has to do with the fact that the psychological impact of terror is so strong, because of the unpredictability. I'm going on an aeroplane: will there be a bomb? That makes people insecure, in a direct way. If there's a war going on, you know where the war zone is, and you know how to keep out of it. Politically, I think most of the terrorists have not actually achieved their goals – their actions have mostly been counter-productive and strengthened their opposition. But indeed, they do have a strong psychological impact.

It may also be argued that in the scenario of a major terrorist attack, with dirty weapons or such, the impact could be tremendous and even exceed what you have in some traditional armed conflicts

Exactly, yes, fear sparks budget increases.

In the study, you measure what you call ‘battle-related deaths’ in a defined area as part of determining the existence of a conflict. While measurement is probably quite feasible in a situation like Israel and Palestine, or in Sri Lanka, it may be much more difficult when you have a global network which is acting. Do you split that up into different situations? Or do you make a global assessment?

We try to do the same assessment for all conflicts, to treat them equally. We do a strict identification of battle events, and see whether we can find the evidence of battles taking place; how many people were involved; how many were killed; who the sides were, and so forth. The main objection is often that there are lots of people dying as a result of secondary effects of the war – the breakdown of the healthcare system, for instance – which is definitely true. But that is not what we include – we look strictly at battle-related casualties. We believe that the health impact is also worth examining; however, it often depends on a number of other factors unrelated to the war, for example the state of the health system before the war.

Do you focus only on the people killed, or also on missing, injured or displaced people?

Yes, we basically focus only on the deaths, because that's often where you have the best information. How many people are wounded will depend on very different understandings of what wounded is, but death is a clear status.

Do you also collect information concerning these other categories of possible victims?

No, not for publication. We may receive that information in various ways, but to go through that systematically – to record how many people were wounded, how many houses destroyed, etc. – would be a huge research project in itself.

Even in that narrow scope of people killed, there are sometimes figures coming out that are quite contradictory, especially regarding the number of people killed in Iraq. How do you measure the numbers, and what sources do you use?

I think that's where much of the debate is right now. These studies are based on epidemiological methodologies: you go in and interview people and try to make them estimate how many people died – then you get these varying estimates. What we do is really try to analyse the battles, the events that take place, and we try to find them by using all kinds of sources. We use a database called ‘Factiva’ quite extensively, which has news information from about 25,000 sources, a lot of information translated from local languages into English, for instance. We also study government reports to congresses or parliaments, amongst others. We try to find researchers writing about situations they have been in, or information they have gathered. Non-governmental organizations also provide very useful information.

We try to get all these versions of sources to establish that these battles actually took place and that there actually were people killed. The epidemiological approach, on the other hand, is often based on what people say happened. I think that can sometimes be a source of exaggeration, as there may not be any inquiry into why people died, whether it really was part of the battle, etc. This is how you get these rather divergent estimates.

So in the case of the two major studies which were done on Iraq, you would tend towards the Iraq Body Count which took a similar approach, rather than the Lancet report which was done based on epidemiological data?

Yes. There is the same debate on the Congo, where researchers – similarly to the Lancet report – also located various spots, went down there to interview people, and then tried to extrapolate from that to the entire nation. In the Congo, it's very difficult, because you don't know how many people there were at the beginning, as the censuses are unreliable. In Iraq the population census might be a bit more reliable. These studies are estimates, more than anything. We just tried to determine where people were killed – you could say that this is maybe underestimating, but over the long term, we have found that our methods work very well. For instance, in Bosnia, from 1991 to 1995, it was constantly estimated that about 250,000 people died. Now, in the big database which has been made by the Research and Documentation Center in Sarajevo, the number is around 100,000, which is much closer to what we reported ourselves. In all these kinds of conflicts, there are reasons for people to want to have very large numbers of victims, but we want to be as accurate as possible.

You use varying outside sources – you don't have your own people in the field measuring?

That's right, we cannot really do field studies ourselves. Over the years, we have learnt to read the material very carefully. You learn, after a while, how many people it is possible to kill with particular weapons, etc. So in that sense, you learn to understand what probably happened.

Do you also take information from humanitarian organizations?

These organisations are very, very crucial, as well as human rights organizations, whose reports are very important sources for us. Definitely, humanitarian organizations have people on the ground who can give very good estimates. Over the last ten years, information on what goes on in various trouble spots around the world has become so much better. Going back in history, it's much harder. Even in the early 1980s, it was difficult to know what happened in Afghanistan, whereas we have a much better account of what happens today. Or think about Indonesia, in the 1950s, or Burma in the 1950s and 1960s. Today we even have more insight into Burma, thanks to various techniques – you can look at satellite imagery, and see whether villages were actually destroyed or not. So I think the world has become somewhat more transparent, in terms of what goes on in organized violence.

The International Committee of the Red Cross is very conservative about giving out figures. Firstly, we don't necessarily have the information, because it's very difficult to estimate the number of people affected by violence, even if you know the situation well. Additionally, we fear that the information we publish may be instrumentalized for political purposes

I agree with a very prudent approach – that's what we try to do. We have an advantage being at the university – the autonomy of the university is very well respected. However, I'm told there are groups and governments who will dispute the figures or definitions. They may have their agendas. We have no other agenda than just reporting these armed conflicts, as comparably and reliably as possible, to provide a basis for research on the causes of conflict or on conflict resolution. The autonomy we have as a university-based organization really saves us. If we were part of the UN, we would be under enormous political pressure. In that sense, we cherish an autonomy that is similar to the ICRC's, and we think that this kind of reputation is what really benefits the world in the long-term.

You mentioned in the beginning that the minimum threshold for categorizing a situation as a conflict was 25 people killed in a year. Isn't that a very low figure in light of the major conflicts going on? In many countries, one could probably come up with examples of inter-communal violence with 25 people killed in a year

It is a very low one, and deliberately so. The tradition was to have a cut-off point at 1000 deaths. Other studies were using 200 or 100, but we wanted to have a low number, in order to capture conflicts when they were fairly small. It works because this makes prevention studies possible, as well as opening the way for new kinds of studies to determine how many of these small conflicts will actually escalate and become big conflicts. Surprisingly few actually do, which is good news. A low threshold also enables us to show that conflicts actually do fluctuate substantially.

Now, of course, as we try to span the whole world, it is not always easy to find information. There are several conflicts where the number of deaths is hovering around the borderline of 25. If we are not sure whether there are 25 dead, or 30, or 20, what do we do? Any cut-off point will have that problem; however, you cannot have just one person killed. For example, the Swedish foreign minister was killed six years ago – that was a one-man thing – but that assassination, we felt, was a different kind of dynamic than what an armed conflict is all about. An armed conflict means that groups are organizing to actually have a political impact, and they're prepared to kill. And if they manage to wage battles with more than 25 people killed during a year, it means that they have a sufficient degree of organization to be counted.

Do you make a distinction, nevertheless, between minor and major conflicts?

We refer to a minor armed conflict when there are more than 25 and less than 999 deaths. From 1000 deaths onwards, we refer to a ‘war’. Clearly, there is a difference, and in the latest report we are doing, we show that there are five such wars in the world, with more than 1000 killed: Sri Lanka, Afghanistan, Iraq, Somalia and Pakistan. When I explain this to people, they are often very surprised – the way the media works, it seems that there are far more than that. But there aren't … of course, ‘wars’ and ‘minor armed conflicts’ is a sub-classification that we only use in our first category of armed conflict, to which we give priority and on which we have the data right now. For this reason, our concept of ‘war’ does not capture the huge number of casualties in Congo, where a lot of fighting will probably be classified under the non-state category.

Do you find that non-state violence fluctuates more?

What we see is that it is more sporadic. It flares up and then is contained. One of the really interesting issues is to ask what happens to these non-state conflicts: why don't they turn into sustained battles? What is being done locally to deal with them? There seems to be a lot of local peace-making going on, which never gets into the headlines, but which is probably very, very significant in managing non-state conflicts.

Would you say it is possible to have a situation where all three categories of violence are going on at once in the same country? For example, in Iraq, where there is a politically driven armed conflict, inter-communal violence, and even acts of one-sided violence?

You can have them in the same place, absolutely. For example, in India, you would have the same: a number of smaller wars going on in North-east India, inter-communal violence in other parts of the country, and the terrorist attacks in Bombay, which are of course in a different category. We think it helps, analytically, to sort them out, to show that they are not necessarily the same and that they would probably need to be dealt with through very different political measures. For armed conflicts, maybe you need political discussion; for non-state conflicts, maybe you need to engage the communal leadership. Terrorism is a separate thing to deal with: how do you deal with suicide bombers, or the organizations supporting them? We think that our classification helps, also politically, in order to get ideas on how to handle the different kinds of conflicts.

You mentioned that presently, there are five situations classified as wars (i.e. an armed conflict with more than 1000 deaths). Do you also have summary figures on the other two categories?

We do our analyses year by year – that figure is for 2008, and we have not yet been able to put together all the data for that year. However, at this point I can indicate the data for 2007 (see map) .

Could you indicate some of the trends you have observed within the different categories, or even across different categories?

There are a number. The first one is that there are now surprisingly very, very few inter-state conflicts. I think that's good news. It means that since the end of the cold war, the incidents of serious international armed conflict have decreased. However, when they happen, they can become very serious, like Ethiopia/Eritrea. You may have had the beginning of such an event in Georgia in 2008, but it was contained very quickly – there was an interest in doing so, as everybody realized that this really could turn into a major interstate conflict.

We have also seen a general reduction in the number of armed conflicts, and that created a lot of headlines. In 1991–1992, there were about 50 armed conflicts, but by 2003, there were 29. Now the number is up to 35 again, so it's escalated – but many of them are smaller conflicts, of which very few have really become big wars. Again, this seems to identify some kind of ability to keep conflicts smaller at least.

What do you believe are the reasons for the reduction in the number of major armed conflicts?

The total numbers now are nowhere near those of the early 1990s. You can attribute that, I think, to increased international activity: the work of the UN, much earlier attention to conflicts, the involvement of additional bodies such as the EU. I also think that the presence of NGOs, who sound the alarms quite early, has an impact.

Another theory on the reduction of conflicts is that democratization and the opening up of societies leads to fewer conflicts actually escalating. I think that's a little too optimistic, but there is a whole debate about democracy and peace. Another argument that has been made is that economic growth in a number of countries has provided the incentive to earn money in other ways than fighting in wars. This strikes me as important. It implies that with this present financial crisis, one should be worried that this may lead to an increase in armed conflict. Maybe that's a sour note to end on.