Lexical Imperfections in the Hague Regulations of 1907

The Hague Regulations (HR) annexed to the 1907 Hague Convention (IV) respecting the Laws and Customs of War on Land were the first reasonably general codification of the law of war (as it was then called) in modern times. Other conventions had codified certain matters (e.g. sea warfare in 1856 or protection of injured or sick military persons in land warfare in 1874). But no codification of the most important matters of warfare from a general perspective had succeeded up to 1899 and 1907.

Nonetheless, the 1907 HR left open at least three problems: 1) the status of the Annex to the Convention (was it binding?); 2) gaps left in regulation (e.g. no general rules on the principle of distinction including the definition of military objectives were included); and 3) lexical imperfections and uncertainties, which invited legal quibbling.

The object of this post is to revisit the third category, lexical imperfections, and to attempt to devise a suitable interpretation of the relevant provisions, particularly considering subsequent law adopted, especially 1977 Additional Protocol I to the Geneva Conventions of 1949 (AP I). It must be conceded that academic literature noted these imperfections immediately after adoption of the Regulations. A prominent example is the (sometimes overly) critical text of Antoine Pillet’s Les Conventions de La Haye in 1918, a little known work in the Anglo-Saxon world.

My analysis will focus on four clusters of problems relating to: the general principle of limitation; prohibited means and methods; bombardment; and occupied territories.

Problem 1: General Principle of Limitation

Article 22 of HR reads, “The rights of belligerents to adopt means of injuring the enemy is not unlimited.” This is the famous principle of limitation. AP I, Article 35(1) reads, “… the right of the Parties to the conflict to choose methods and means of warfare is not unlimited.” Here, there is a series of defects, some minor, others more significant.

To speak of “rights” is a pardonable imprecision. While belligerents enjoy some subjective rights in war, in the context of the conduct of hostilities most belligerent legal positions are “faculties” rather than “rights.” A subjective right includes an obligation by another party to do something or to suffer something. This is not the case in the conduct of hostilities. A belligerent can do everything in its power to ward off action by an opposing party. It has no obligation to accept hostile actions with passivity or to suffer these acts, what in Latin is called pati. However, because the vocabulary of “rights” is by now well-established—as it is in UN Charter, Article 51 for self-defense (which is also a faculty rather than a right)—the issue can be left there.

A more annoying imperfection is the mention of “means” alone in the prohibitive rule, an imperfection that AP I remedied. Taken literally, the rule conveys that only “means” (i.e. tangible devices, such as weapons) are subjected to the principle of limitation and not “methods” (i.e. intangible devices, tactics). This is nonsensical. One way to read “means” is by folding in “methods” to the meaning of the word. Another correct way to read the rule is to use “means” as a gap filler that includes “methods” by analogy. That this is the correct meaning is made abundantly clear by the later law codified in AP I.

Furthermore, the phrase “injuring the enemy,” suppressed in 1977, is equally imperfect. It could give the impression that the principle of limitation includes only acts which have some physical effect on the adverse party. This, however, is not a correct interpretation. A means or method is prohibited because of the danger it causes or the inadequacy it carries, not only because of concrete injuries it causes. In this sense, using a flag with protective emblems to transport weapons may not “injure” the enemy, unless the term injury is correctly interpreted as a purely abstract link to all acts, omissions, and facts that are contrary to the law of armed conflict. Once again, AP I remedied this lexical imprecision by eliminating the relevant part of phrase.

Problem 2: Means and Methods Prohibited

Article 23(a) of HR, includes the prohibitive text, “To employ poison or poisoned weapons.” This text creates doubts about gaseous forms of poisons. It has been said that gases were covered by the Geneva Gas Protocol of 1925 precisely because that provision in the HR does not extend to poison other than in liquid or solid form. This is not the interpretation given today to this provision, which is extended to poisons in the form of gas. Debate over the scope of Article 23(a) could have been avoided by more careful drafting (e.g. by using the term “poison, in all its forms,” or by adding “in solid, liquid or gaseous, or any other form”).

Subsection (b) of the same provision forbids, “To kill or wound treacherously individuals belonging to the hostile nation or army.” It could be said that treacherous killing is always prohibited, not only against enemy nationals. But that is a matter of scope of application of the law of war, which does not extend to the treatment of its nationals by a State. Another aspect is graver; it stands to reason that members of the adverse army shall not be killed or injured by treachery (what today is called perfidy). But what does that mean with regard to individuals belonging to the hostile nation, i.e. non-combatants or civilians?

The rule here is certainly underinclusive. If it means that civilians shall not be killed treacherously, this remains well within the principle of distinction that civilians shall not be directly attacked in the first place (unless they participate in hostile action, under modern vocabulary, “direct participation in hostilities”). The correct interpretation is thus not a contrario (non-treacherous killing is not prohibited), but treatment by analogy with the underlying principle of Articles 25-27 of HR, the protection of the civilian population, and today with AP I, Articles 48 et seq.

Subsection f of the same HR provision forbids, “To make improper use of a flag of truce, of the national flag or of the military insignia and uniform of the enemy … .” This drafting is overinclusive. At the time of its adoption, it was unclear whether the wearing of an enemy uniform was generally prohibited or only so when used to interfere with military operations. Practice was split after the Second World War, but the rule linking the prohibition to military operations clearly prevailed and can be found today in AP I, Article 39(2). The rule in the HR must thus be interpreted in this restrictive sense, by adding an element which is not there. It can also be noted that the conjunction “and” between insignia and uniform should read as being a disjunctive “or,” because it is sufficient to abuse either insignia or uniforms, as is made clear now in AP I.

Subsection (g) of Article 23 raises the question whether only “destruction” and “seizure” of property are thus prohibited, but not other forms of interference. By way of comparison, one can recall the terms used in AP I, Article 52(2) (i.e. destruction, capture or neutralization). Capture will correspond to seizure, but neutralization, which conveys the idea of rendering inoperable, is not mentioned in the HR. Under modern conditions, it should be interpreted into the text, probably under an enlarged conception of “seizure” (as in the case of expropriation in another area of international law: formal and material expropriation, here formal and material seizure).

Subsection (h) of Article 23 of HR prohibits, amongst other things “to compel the nationals of the hostile party to take part in the operations of war directed against their own country.” The criterion is not overly clear. The true criterion seems to be that all services that augment the power of the adverse army shall not be compelled. This is more than mere participation in the “operations of war” which conveys the idea of a participation in the hostilities. The expression used in the HR does not seem very precise. The new law is now set out in Article 50 of the Third Geneva Convention of 1949, taking great pains to prohibit any contribution of military nature, even for public utility services.

Problem 3: Attacks and Bombardment

The rules on bombardments (Articles 25-27 HR) are quite poorly drafted. Perhaps they were the best one could get in 1899 and 1907, but they have hardly stood the test of time.

Article 25 HR reads, “The attack or bombardment, by whatever means, of towns, villages, dwellings, or buildings which are undefended is prohibited.” This is the famous “open town” provision, replaced in the modern law by the more sweeping “non-defended localities” provision under API, Article 59.

Bombardment was a new term at the time. It was added to Article 25 for clarity. But because a bombardment is an attack (of massive nature), the term does not add anything legally useful to the general term “attack,” which is the controlling one. The juxtaposition of attack and bombardment could, however, convey the false idea that the only offensive attacks by besieging forces are at stake. This is so because bombardments in due form are and were performed by “besieging” offensive forces. But the correct interpretation must be that all forms of attack, offensive and defensive (e.g. when the adverse forces are already in the town) must conform to the rule. This is now made clear by the better formulated text of Article 49 AP I.

The term “undefended” refers to a declaration of non-defense. The term must thus be interpreted in accordance with military practice; the implicit reference to such practice is not a lexical defect. Equally, the objects mentioned must be understood as only illustrative and not exhaustive.

The separate mention of dwellings is also strange; it is not ejusdem generis with regard to the other three objects mentioned, which are buildings and series of buildings. Consequently, it would have been better not to mention it to avoid giving rise to some misunderstanding as to the true class of objects covered. The best formulation would be to manifestly refer to civilian objects. Today, Article 25 of HR should be interpreted to refer to all such objects, when considering subsequent practice and later conventions related to it, especially AP I (under the 1969 Vienna Convention on the Law of Treaties, art. 31(3)(b) and (c)). We must however keenly note that in the formulation of the HR, an attack on civilian objects seems permitted a contrario when the town or locality is not “undefended.” This is most unwelcome.

In Article 26 of HR, there is a strange misalignment regarding Article 25. Apparently the warning requirement applies only to “bombardments” there mentioned, not to “attacks,” a word which is not replicated here. This omission is faulty; attacks must be interpreted into the provision. There should be no quibbling as to when an attack has the intensity of a bombardment. This is true even more because the protective aim of the provision applies in all attacks, not only to the most intense ones.

Article 27 of HR also presents several difficulties. That article reads,

In sieges and bombardments all necessary steps must be taken to spare, as far as possible, buildings dedicated to religion, art, science, or charitable purposes, historic monuments, hospitals, and places where the sick and wounded are collected, provided they are not being used at the time for military purposes. It is the duty of the besieged to indicate the presence of such buildings or places by distinctive and visible signs, which shall be notified to the enemy beforehand.

First, the objects mentioned here are, in modern parlance, a strange mixture of objects under general protection and objects under special protection. Hence, it designates a narrower category than the total sum of civilian objects which should have been protected under Article 25. Some of the objects under general protection are merely civilian objects (such as buildings dedicated to science), other objects are under special or enhanced protection (such as historical monuments). In the latter case the distinctive signs (notably protective emblems) make good sense, but not in the former case. This disordered conflation of objects is more than a lexical failure; it is a conceptual pitfall.

Second, use for “military purposes” is too narrow a category. A historical building can be used for minor military purposes, such as to route communication wiring, but the military advantage of destroying the building may appear too limited to warrant an attack. The reference to “military purposes” is also vague. A contribution to military action is tangible, the reference to purposes is not. After all, collecting money may fulfill military purposes.

Hence, the law under Article 27 must be interpreted today in light of AP I, Article 52(2), in the double sense of referring to military contributions and by adding the criterion of military advantage.

Finally, protective signs are regulated in a way which does not correspond to practice. There is no duty to use such “emblems” and protection does not depend on them. Modern law, such as AP I, Article 56(7) clarifies this. Because protective signs at that time were not yet uniform, and they were to be applied also to objects under “general protection,” the duty of notification to the enemy was understandable. That former shortcoming of the law of war is today overcome.

Could such gaps in the regulation of bombardments be mitigated by reference to the Martens Clause? That depends once more on an act of interpretation. It is possible to argue that the cases we identified (e.g. bombardment of civilian buildings) are not clearly regulated in the text and thus the Clause could residually apply. However, the Clause is meant to apply to cases “not included” in the Regulations. It is possible to say that bombardments were precisely included, and that apparent “gaps” must thus be interpreted a contrario as meaning freedoms of action (Hague Convention IV, 1907, preamble, § 8).

Problem 4: Occupied Territories

The regulation of occupied territory in Articles 42-56 of the HR gives rise to a series of lexical problems, of which only some can be discussed here. Various problems of interpretation of the provisions will also not be raised. I will focus instead on lexical imperfections in the formulation of these provisions. One may notice that in general—as seen in the Fourth Geneva Convention of 1949 (GC IV)—the distinction between invasion (and the duties applicable there) and occupation (and its applicable duties) has not been made.

This omission may pose significant problems of interpretation because some duties formally applicable in an occupation phase may also be called to apply to the invasion phase (today, for example, Article 49 of the Fourth Geneva Convention (GC IV) on deportations and internal displacements; and Article 47 of HR on pillage, or Article 52 of HR on requisitions). This has triggered discussions on a “functional application” of the law of occupation in pre- and post-occupation phases. Moreover, most of the provisions of the section on belligerent occupation are drafted in general terms and leave open many questions, giving disproportionate room to differing interpretations.

Article 43 HR holds, “The authority of the legitimate power having in fact passed into the hands of the occupant, the latter shall take all the measures in his power to restore, and ensure, as far as possible, public order and safety, while respecting, unless absolutely prevented, the laws in force in the country.” This formula has been criticized for the use of the words “in fact.” These words refer to one of the great controversies of the beginning of the 20th century, i.e. the extent to which belligerent occupation is a power of fact or seen as a power in law.

In terms of power of fact, some States did not want to recognize any rights of the adverse belligerent on their territory, that being for them a high treason or an admission of defeat. In parallel, if occupation is a power in law, the occupation of a State triggers a series of rights and duties of the occupant and occupied State under international law. Article 43 of HR seems to take a position aligned with the “factual” school.

However, there are various rights, and especially duties, of the occupant set out in the provisions that follow the rule. In that sense, belligerent occupation is not only a situation of fact, but also a legal state of affairs, a situation triggering a series of subjective legal positions (rights, duties, faculties, etc.). The words “in fact” must thus be interpreted as meaning that the fact mentioned also triggers rights and duties. It is what in French is called an acte déclencheur. If belligerent occupation is also a situation of law, the acts of the occupant, to the extent they are compatible with the empowerments under the law of occupation, are opposable to and bind the occupied State. This is true especially after the return of the occupied territory under its power, corresponding to international case law, but not reflected in the “in fact” formula of Article 43.

Moreover, the mentioned provision asks the occupant only to ensure “public order and safety.” It is interesting to note that no mention is made of all the humanitarian supplies necessary for the survival of the civilian population in occupied territories, as is now done in Articles 55 and 59 of GC IV. Consequently, from the perspective of today, the terms “public order and safety” should be interpreted as covering the necessary humanitarian supplies, as an indispensable element of public “safety.”

It stands to reason that such post festum interpretations are not truly necessary now because it is sufficient to apply the later law, be it conventional or customary. However, the HR must be interpreted themselves as having moved by evolutionary interpretation to be aligned with the modern law, as seen in the notion of “war” contained in the Hague codification. Today, “war” is aligned with the notion of “armed conflict” stipulated in common Article 2 of the Geneva Conventions of 1949.

Finally, we may note that the “laws in force” which were not to be modified, as far as possible, were essentially meant to be private laws. For public legislation, the principle was rather the reverse; not only could the occupier suspend such laws for absolute military necessity, but for all types of incompatibilities with its policies in the territory. Thus, the formula as it stands is better suited for private law (where the exception is truly necessity) than for public legislation (where the exception is considerably broader).

Article 46 of HR reads, “Family honour and rights, the lives of persons, and private property, as well as religious convictions and practice, must be respected. Private property cannot be confiscated.” This provision includes unwelcome gaps which must be filled by interpretation. In particular, the freedom of persons (habeas corpus) must be included in the definition of “their lives.” It is strange that collective “rights” (in the context of families) were mentioned but not individual rights pertaining to persons.

It can be inferred that the right to freedom is not unqualified in occupied territories. It is limited by curfew or by the possibility to detain persons considered dangerous by the occupying power, e.g. for sabotage or spying. By the same token, it is strange that the prohibition to take hostages has not been voiced, contrary to the modern law (see e.g. Article 3 common to the Geneva Conventions or Article 33 of GC IV). After all, the practice of taking hostages was already employed by Germany in the 1870 Franco-Prussian War; it was therefore fully present in the minds of the delegates. Overall, Article 46 is drafted in a very thin manner and is full of gaps.

Article 48 of HR reads,

If, in the territory occupied, the occupant collects the taxes, dues, and tolls imposed for the benefit of the State, he shall do so, as far as is possible, in accordance with the rules of assessment and incidence in force, and shall in consequence be bound to defray the expenses of the administration of the occupied territory to the same extent as the legitimate Government was so bound.

Again, this drafting is highly imprecise. On the face of that text an occupier may levy taxes for many purposes, even possibly for the benefit of its own population. Indeed, nowhere is it said that the taxes must be used to the benefit of the occupied territory. Moreover, the comparison with the administration of the legitimate government is odd. The factual and legal bases for the administration are not the same; the legitimate government administers as a sovereign and with regard to all public functions, while the occupier administers in the limited context of belligerent occupation. The expenses of the administration are thus difficult to compare, something which the provision manifestly suggests in the last part of its sentence.

Finally, Article 50 reads, “No general penalty, pecuniary or otherwise, shall be inflicted upon the population on account of the acts of individuals for which they cannot be regarded as jointly and severally responsible.” This formulation is also ambiguous. The last part of the sentence seems to have been inserted as a descriptive element, giving the reason for the rule, i.e. that the rule is there because the population in general is not responsible for such acts.

However, the reasons for a rule are never inserted into its text. Thus, in our case, the drafting suggests that some individual acts exist for which the population at large can be held jointly and severally responsible. This is so because the exclusion operated only “on account of the acts of individuals for which they cannot be regarded … .” This language implies a contrario that such responsibility may exist for acts of individuals which can be so regarded. It might be that such an attribution is reasonable for some groups of persons, but it will not be admissible for the “population” at large. That, however, is precisely what the text suggests. The inconvenience of drafting is manifest. It opens the path for arbitrary interpretations and collective punishments.

Conclusions

At this juncture, we can conclude our analysis of the various provisions of the Regulations and can conclude several things. First, the drafting of the HR is often sobering in its brevity and leaves many gaps. Second, the HR contain a series of lexical defects which give rise to uncertainties as much in bona and in mala fide interpretation. Third, it is possible today to fill these gaps and to settle the interpretation by having recourse to the later law, related to the HR by the tools of subsequent practice and contextual interpretation. Last, the reasons for the state of the 1907 codifications, so different from the following ones, is the decidedly 19th century perspective taken. Under this belief, a detailed and neat regulation is useless because either a belligerent is civilized and bona fide, and all difficulties will disappear, or a belligerent is uncivilized and mala fide, and no text will suffice. It is essentially this last theorem which has been cast off since 1907, in the shift from the old world of the laws of war to the newer world of the law of armed conflict and later that of international humanitarian law.

***

Robert Kolb is a Professor of Public International Law at the University of Geneva and former legal staff member of the ICRC. Prof Kolb is also a member of the legal section of the Swiss military high command (IHL section).

The views expressed are those of the author, and do not necessarily reflect the official position of the United States Military Academy, Department of the Army, or Department of Defense.

Articles of War is a forum for professionals to share opinions and cultivate ideas. Articles of War does not screen articles to fit a particular editorial agenda, nor endorse or advocate material that is published. Authorship does not indicate affiliation with Articles of War, the Lieber Institute, or the United States Military Academy West Point.