From 19th century until soon after WWII, major disputes between the countries would typically see a quick de jure settlement once the hostilities (if any) came to a halt and a certain de facto status quo was achieved:

  • Wars would end with peace treaties and restoration of diplomatic relations. After Franco-Prussian war, a peace treaty has been signed in four months, in which France officially and in written recognised Alsace-Moselle as German territory, and restored diplomatic relations. Thus, although there were revanchist sentiments, there were no formal territorial disputes between the countries. Likewise, both Crimean war and Russo-Japanese war were concluded with a peace treaties in less than 4 months after the armistice;
  • De-facto independent new states (or new government) were recognised by all or almost all other states faster than it happens nowadays. For instance, Greek war for independence ended in 1830, and as soon as in 1832, the independence of Greece was recognised by the Ottoman Empire. Similarly, Romania and Montenegro achieved fully recognised independence in 1878, nearly at the same time as they achieved de facto independence. Serbia might be a counterexample, as it declared independence (and achieved it de facto) in 1869 and it only was recognised in 1878. Brazil declared independence in 1822 and it was recognised by Portugal in 1825. With Spanish American colonies, it took a bit longer, from end of hostilities in 1829 to recognition in 1836 to full diplomatic relations by mid-1840-s. Of course, following WWI, many new countries appeared, and most were relatively quickly recognised.

After WWII and the collapse of colonial empires, the pattern has changed considerably:

  • Wars often don't end with peace treaties, and there are territories that remain formally disputed for decades. For example, after WWII, it took 11 years to formally end the state of war and a proper peace treaty between Japan and USSR or Russia has never been signed. For last 73 years, Japan disputes the status of Kuril islands. The two Koreas officially claim each other's territories 65 years after the war ended. Israel holds some of the territories it occupied in 1967 to this day (for 51 years), with status far from being settled.
  • There are many de-facto independent, but largely unrecognized states. Taiwan has been independent for 69 years, but is still only partially recognised. The same applies to Northern Cyprus (45 years of de facto independence); Somaliland, Karabakh, Abkhazia, South Ossetia and Transnistria (some 25-30 years); Kosovo (10 years); more recently, east region of Ukraine.

Is this observation valid? Is it true that post-WWII, it has become more common that a de facto state of affairs, e. g., control over territory or independence, lacks formal recognition for a long period of time? Is it possible to quantify this 'more common' claim?

What are some reasons for this? Are there less incentives to reach formal agreement now? Does persistence give more leverage? Do the reasons have to do with international affairs or with domestic politics (e. g., the need to convince the electorate)?

  • @Kostya_I: You have also completely ignored the numerous wars of independence in the Balkans through the entire 19th century, right up to 1910. Jul 23, 2018 at 17:51
  • Please do not reply in comments; comments should request clarification; clarifications should be edited into the question.
    – Mark C. Wallace
    Jul 23, 2018 at 22:28
  • 1
    I have attempted a friendly edit; is more work needed, or can the question be nominated for re-opening? If more work is needed, who can help? Can some of the examples in the comments be moved into the question to clarify?
    – Mark C. Wallace
    Jul 23, 2018 at 22:37
  • @Kostya_l Why do you mention an unrecognized Basque government? I never heard that the ETA had formed some type of shadow government.
    – MAGolding
    Jul 24, 2018 at 3:25
  • 1
    I don't include Basque (it was included by other editors). ETA has not achieved control of their territory, so there's no discrepancy between de facto and de jure here.
    – Kostya_I
    Jul 27, 2018 at 1:02

2 Answers 2


Maybe the root of your question aim to the surge of the concept of nationalism, and later on self-determination and the United Nations.
Nationalism surges as a concept after French Revolution, and later on in 1848, it means that countries should be composed by a single nation of people. While self-determination surges with United Nations, where its idea is that people should select how they want to chose their sovereignty.
In the past, a territory was an asset of a king or a country, so they had the right to yield a territory as an exchange of peace. But with the new concept of self-determination, it was no longer acceptable to give or take territory without consulting the people that lives there.
With the current concept of self-determination, people of Alsace and Lorraine should have been consulted whether they wanted to be germans of frenchs.

Based on these concepts, is no longer easy to set peace terms, because any exchange of territory involves the right of the people of that territory to revolt against their new ruler (for example, the West Bank territories controlled by Israel). Or the contrary, a nation of people can demand their own country agaist their current ruler (Kosovo).

Besides, United Nations concept is to prevent wars between countries, not to solve problems of self-determination. Hence, war is no longer acceptable to solve territorial disputes like it was in the past.

  • the right of the people of that territory to revolt against their new ruler (western bank controlled by Israel) -> I don't understand that bracket ?!
    – Evargalo
    Aug 2, 2018 at 8:12
  • 2
    As a complement, the concept of "international law" protected by the UN and forbiding wars of conquest makes a written yield of territory much more definitive. While France loss in 1871 lead to the loss of Alsace-Lorraine "until next war", if today, say, Ukraine was to officially recognize Crimea as Russian they would lose any legitimacy for a later reconquest (be it by a war, a referendum, a revolt of Crimean...) and it would legitimize in retrospect the Russian aggresion (probably ending western sanctions against Russia, for instance)
    – Evargalo
    Aug 2, 2018 at 8:18
  • Hi @Evargalo. Regarding your first comment, I'm talking about palestine population living under israeli rule, that people was not consulted if they wanted Israel as their goverment. About your second comment, maybe you're right about Crimea. But Donets and Luhansk might be an issue more complex to solve.
    – Santiago
    Aug 2, 2018 at 13:14
  • Oh yeah, Western Bank in Palestine. I was absent-minded and wondering if this was some complot theory about Israel ruling banks in the Western hemisphere or smg like that ! :)
    – Evargalo
    Aug 2, 2018 at 13:25

If we stick to territorial disputes, for which there exists a historical database (ICOW = Issue Correlates of War), which however does not include secessionist claims unless they are part of an annexation claim (so the Catalans for example don't qualify), then some things can be said, but it's not a definitive answer.

The main thing that can be said is that fewer waring countries apparently correlate with fewer territorial disputes getting settled... because:

At first glance, relatively few territorial claims were settled directly through military conquest. However, this statistic is somewhat misleading, as many territorial claims settled through other methods -- including particularly bilateral negotiations and peace conferences such as Versailles in 1919 -- reflected battlefield results. If we recode settlements that arise from military outcomes as being settled by military conquest, the patterns change considerably. The important role of military conquest in settling territorial claims becomes clear, ending roughly one in four claims, and trailing only bilateral negotiations in frequency.

Of course this doesn't quite answer your question because bilateral negotiations may work faster in the present. I have no idea if that's the case, since I can't find any published results on that.

I suspect one could use the ICOW database to answer most of your question directly (e.g. compute the time difference between settlement and first claim) and plot that vs (begging or [probably better] end of claim) time. I'm not aware of a paper that has done that already... and my quick attempt to do it turned up the obvious problem: open claims (which don't end when the database does) bias the result heavily towards the present being faster if the open claims are excluded. One probably needs to do something like survival analysis...

And here's my quick and dirty survival plot

enter image description here

So it looks like claims ended (not just settled, because it might have been just abandoned to be counted as ended) after 1950 took longer to be... ended. As one can guess from the Kaplan–Meier plot, the difference is statistically significant (p << 0.001):

survdiff(formula = claim_surv ~ ends_after1950, data = claimdy_df)

                   N Observed Expected (O-E)^2/E (O-E)^2/V
ends_after1950=0 169      169      121      18.7      31.8
ends_after1950=1 247      157      205      11.1      31.8

 Chisq= 31.8  on 1 degrees of freedom, p= 1.73e-08 

Some disclaimers:

  • I don't have a lot of experience doing survival analyses; this post is nowhere near the level of rigor of an academic science publication
  • The data may be very heterogeneous in other ways; in particular there's things like river and sea fishing rights disputes in the database. If I have more time to spend on this, I'll try to see if the conclusion is robust by excluding non-land claims and/or considering how the dispute got ended.

  • Also, the 1950 cutoff was chosen fairly arbitrarily by me to approximate "after WWII". The result might be sensitive to the cutoff as well.

  • All the right-censored (meaning still open) claims (that's what the "+" signs are in the K-M plot, by the way) are obviously gonna go in the "after-1950" bin; this creates a problem for the open claims that lasted longer than 50 years (database ends in 2001); these >50-years-long and still open claims had a chance to be settled both before and after 1950. But in my quick analysis I'm only counting them as unsolved after 1950 (they were also unsolved in a sense before 1950 as well).

I also threw a Cox proportional hazards regression at the same simple model (but using the end date of the claim as a continuous covariate, i.e. no binning by some arbitrary cutoff year), even though I expected the proportional hazards assumption to be violated; the proportional hazards assumption here is that for every unit of time [month] increase in the end date the conflict will have lasted proportionally less, and indeed this assumption is seriously violated

            rho chisq        p
endclaim -0.191  10.9 0.000956

My main reason for trying Cox proportional hazards was to see the Schoenfeld residuals (graph below, time in years on X axis is conflict duration). These residuals suggest that instead of the flat line one would expect for proportional hazards, there is downward trend, meaning that as conflicts last longer (are more intractable), it makes less of a difference what the end date was as a factor for their duration; in other words, more protracted (longer) conflicts appear less influenced in their duration by when they get solved (i.e. by their end date).

enter image description here

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .