2023 DoD Manual Revision – A Commentary on the Amendments


| Aug 23, 2023


(Editor’s note: This post is part of a series analyzing the 2023 revisions to the U.S. Department of Defense’s Law of War Manual.)

Introductory Remarks

Shortly after the 2015 publication of the US Department of Defense (DoD) Law of War Manual, Professor Michael Newton kindly invited me to contribute a chapter to the book he was planning in which he would present a multi-author assessment of the Manual. I accepted his generous invitation, and having written the chapter, I continued to prepare my own assessment of the DoD text. It rapidly became apparent that putting together a thorough commentary on the approximately 1,200 pages of the Manual was going to be a considerable task, and Professor Wolff Heintschel von Heinegg agreed to share the load with me. The result, The Law of War – A Detailed Assessment of the US Department of Defense Law of War Manual, was published in 2018. In May and December 2016, two sets of amendments to the Manual were published and most of the substantial amendments were considered in the final chapter of our book. Now, in 2023, some significant further amendments have been issued. This post updates The Law of War text by scrutinising those more recent amendments and by offering my own assessment.

While it is not necessary to trace in any detail the history of the evolution of the relevant text, it is sensible to point out that a number of the topics addressed in some detail in the following pages were dealt with rather differently in paragraph 5.5 of the original, 2015 Manual text. Specifically, sub-paragraph started by quoting the doubt rules in Articles 52(3) and 50(1) of Additional Protocol I (AP I). It then asserted, “Under customary international law, no legal presumption of civilian status exists for persons or objects, nor is there any rule inhibiting commanders or other military personnel from acting based on the information available to him or her in doubtful cases.” Later in the same sub-paragraph there is the following statement, “A legal presumption of civilian status in cases of doubt may demand a degree of certainty that would not account for the realities of war. Affording such a presumption could also encourage a defender to ignore its obligation to separate military objectives from civilians and civilian objects.” It is then noted that “unprivileged belligerents may seek to take advantage of a legal presumption of civilian status.” As Professor Heintschel von Heinegg and I noted in our book, the U.S. objections to the two AP I doubt rules were already well known, so further comment in that regard when preparing our book in 2016 and 2017 seemed inappropriate.

The changes to the Manual text promulgated on 13 December 2016 moved the text discussed above from paragraph 5.5 to 5.4 and made further minor adjustments. The assertion that the two AP I doubt rules do not have customary status remained, however, and the departure from that position is the most striking feature of the 2023 amendments that are discussed below.

This post examines the Manual text as now amended. My comments are intended to be constructive and recognise the considerable effort required to put together, and then maintain and update, a manual of this size. In particular, I recognise that the Manual text, as amended, is intended to reflect the U. S. Department of Defense position on these matters, a position which is not necessarily shared by all elements of the U.S. government, nor indeed by all commentators. Nevertheless, the U.S. DoD is to be congratulated on the continuing commitment it has demonstrated to update the text from time to time.

Opening Text of Paragraph 5.4

The 2023 amended wording of paragraph 5.4 starts with a title, “Assessing Information in Conducting Attacks,” which suitably reflects the ensuing content. The opening lines of the paragraph then address “Persons who plan, authorize or make other decisions in conducting attacks,” wording which usefully teases out the persons who have the obligations that are then discussed. While the word “reasonably” is still not used to qualify the notion of available information, the text does usefully spell out by way of example that the commander must, on the basis of available information, determine in good faith that a target is a military objective before authorising an attack against it. The amended text also points out that “the expected incidental damage to civilians or civilian objects must be assessed in good faith given the information available to the commander at the time.” As a counsel of perfection, it might have been preferable in this part of the text to refer again in sentences two and three to “persons who plan, authorize, or make other decisions in conducting attacks” if only to avoid any impression that the two later sentences only apply to commanders. However, what is good about the recent amendment to this part of the text is that it is now made clear that the commander must determine in good faith that the object of the attack is a military objective before attacking it. Moreover, the reader must go on to consider carefully what is written in the remainder of paragraph 5.4.3.


There is, perhaps, a slight awkwardness in the drafting of sub-paragraph In my view, international law requires the attacking decision-maker to consider all relevant and reasonably available information in determining whether the target is a military objective and whether the indiscriminate attacks prohibition would be breached. The Manual states that in making these decisions, “individuals may rely on the information that is available to them, whether they have personally observed it or obtained it from other persons or means.” It may be that the Manual’s authors are distinguishing the notion of considering information from the idea of relying on it. If so, perhaps this should be made plainer. It might, for example, have been clearer if the sub-paragraph had begun with the point made in the second sentence of the current paragraph, following this with something along the lines of “Inevitably some pieces of information will be found to be more reliable and useful than others.” The sub-paragraph might then have continued “Having considered all relevant and available information, the decision maker should rely on the information that he or she finds to be the most reliable and useful, whether that derives from personal observation by the decision-maker or from other persons or sources.” The distinction between considering information and relying on it is, one would suggest, an important one. The reference in footnote 87 is useful, noting of course that the observations were being made in the context of the potential application of international criminal responsibility as opposed to a determination whether the law of armed conflict obligation had been complied with.


Sub-paragraph essentially represents new text. The topic it covers is amply set forth in the title, “Classifying Persons or Objects as Military Objectives when Planning and Conducting Attacks.” This is, of course, something that lies at the very root of lawful targeting decision-making, so it is good that it is now being explicitly addressed in the Manual in this way. In the first sentence, the limitation of attacks to military objectives is properly stated, and the existence of other rules to protect civilians, other protected persons, and objects is noted. The second sentence then explains, properly, that when deciding on or planning attacks the decision whether a person or object is a military objective must be made in good faith based on available information.

The next sentence reads,

In addition, these decisions must be consistent with the obligation to take feasible precautions to verify that the objects of attack are military objectives and with other obligations to seek to reduce the risk of incidental harm to civilians and other persons and objects protected from being made the object of attack.

One wonders whether this sentence might have been more clearly drafted. The “these decisions” must be referring to the decision whether a person or object is a military objective, and thus a lawful object of an attack. The relevant international law obligation, here, is to take all feasible (that is, practicable or practically possible) steps to verify that the object or person in question is, indeed, a military objective. The other precautions prescribed by Article 57 of AP I and customary law are to be considered if, and only if, the decision is made that the person or object is indeed a military objective. Perhaps, therefore, planners and decision-makers would better understand the legal requirement if the sentence were to read, “In making the decisions referred to in the previous sentence, all feasible, that is all practicable or practically possible, precautions must be taken by attack planners and those deciding upon attacks to verify that the person or object to be targeted is, indeed, a military objective.” Precautions to limit collateral damage only arise after the person or object has been properly classified as a military objective, and it is not therefore appropriate to refer to that aspect in this part of the text.

The next part of sub-paragraph is as follows, “Under the principle of distinction, commanders and other decision-makers must presume that persons or objects are protected from being made the object of attack unless the information available at the time indicates that the persons or objects are military objectives.” This could be misunderstood as seeking to combine what are, in law, two distinct rules, namely the doubt rules and the rules as to the definition of civilians and civilian objects. It is the use of the word “presume” that causes the potential ambiguity here.

If the Manual’s authors intended that this sentence should actually address the second of these topics, it would have been better, perhaps, if the sentence had read, “Under the principle of distinction, persons or objects that are not military objectives are, respectively, civilians or civilian objects, and must not therefore be made the object of attack.” If, however, the text was intended to refer to the doubt rules, the effect of the current formulation, as Hitoshi Nasu and Sean Watts have persuasively pointed out, is that rather than articulating “a means to resolve doubt, the Manual’s presumption almost seems a universal ‘going in’ proposition. It instructs soldiers to apprehend their environment as entirely civilian.” The analysis of those two commentators indicates that the Manual’s formulation here restricts attack options significantly more than the AP I doubt rules, not least because the latter only arise if doubt exists and, in the case of the UK for example, if that doubt is significant. Consider that Professors Nasu and Watts’s post notes that the presumption is only triggered when there are missed indications that create doubt about targetability and where the interpretation by the UK that substantial doubt must exist for the presumption to arise further restricts its applicability. Nasu and Watts then consider whether the presumptions are in fact customary, helpfully noting the HPCR Air and Missile Warfare Law Manual’s formulation in Rule 12b.

The next sentence then talks about the presumption as being the starting point for the good faith exercise of military judgment based on information that is available. These are potentially opaque ideas, and one cannot help thinking that a simpler formulation along the lines discussed above would say all that really needs to be said at this stage. The next three sentences are, in my view, correct and do not require further discussion.

Good Faith

Sub-paragraph continues under the sub-heading of “Good faith.” At the risk of being pedantic, the first sentence should arguably, for grammatical and clarity reasons, read “Commanders and other decision-makers must assess in good faith whether persons or objects are military objectives.” The second sentence puts the point well, i.e. commanders and other decision-makers, the latter notion including for present purposes operational planners, must indeed have an honest and genuine belief that the person or object to be attacked is a military objective. The subsequent sentence is also right in noting that the law of war does not require the application of a fixed standard of evidence or proof. Indeed, it might have been useful if this part of the narrative had reminded the reader about the inappropriateness of employing litigation-based notions such as the concepts of evidence and proof to targeting decision-making. This could, perhaps, have been achieved in part by cross-referring to footnote 92. It could usefully have been pointed out that such decisions are, in practice, based on information which might well be in a form that would be inadmissible before a judicial process but that is nevertheless a suitable basis for reliable decision-making. Referring in the sentence to information would also link the discussion in the first “Good faith” paragraph to that in the text that precedes it. The extract from the ICRC Interpretive Guidance in footnote 98 is generally helpful, but my point is that the main text of the Manual might usefully have made the suggested points.

The reference to “professional judgment” and to reasonableness in the “Rather” sentence is helpful and, it is suggested, accurately reflects lawful practice. The reader might then wonder about the sentence starting with the word “Relevant.” In this sentence the reader is being told what circumstances are relevant when deciding what action is reasonable in making the military objective determination; this, at least, seems to this author to be what the “Relevant” sentence is referring to. The first example of a relevant circumstance is said to be “the time and resources reasonably available.” It must indeed be the case that the time that is available will influence what kind of action can reasonably be taken to decide that a person or object can lawfully be targeted. By way of peripheral comment, it is undoubtedly the case that the time available for attack decision-making in some situations may be very short. The development of certain stealth, hypersonic, autonomous, and cyber weapons should be considered in this context. Nevertheless, the way in which weapons to counter such threats are employed must be such that a determination, whether human or otherwise, that the object of the countering use of force is a military objective can properly be made. The degree of risk that the attack will pose to civilians is also a relevant consideration, as is the risk to friendly forces and the military advantage anticipated, rather than expected, from the attack and it is good that these aspects are also mentioned.

The important point that attacks must not be directed against persons or objects based on mere speculation is well made, and the kind of “signature strike” attack decision-making that would be regarded as speculative, and thus unacceptable, is explained clearly and succinctly, usefully supported by references to the Galić decision and to the New Zealand Manual in footnotes 100 and 101.

Available Information

While there is an inevitable degree of circularity, the opening sentence under the “Based on the Available Information” heading articulates the decision-maker’s obligation to determine military objective status based on available and relevant information. This is the key point and is clearly stated. Arguably, the remainder of the narrative under this heading would have been a useful place in which to identify some of the potential sources of such information, including reconnaissance assets, espionage, material gained from the questioning of prisoners, intercepted enemy communications, observed enemy military movements, and so on. Having set forth the potential sources of such information, the paragraph might usefully have teased out the aspects of the required decision-making processes that the information should support, such as the determination whether persons are combatants, non-combatants, directly participating civilians or peaceable civilians (cross-referring to where those matters are respectively discussed elsewhere) or whether an object satisfies the definition of a military objective (again cross-referring to the discussion of that definition). One wonders whether that approach might prove more helpful to the reader confronted with making such decisions. A useful point is made in the extract quoted in footnote 103.

Feasible Precautions to Verify

The first sentence under the “Feasible Precautions to Verify” heading is clearly correct and does not require further discussion. In the next sentence, feasible precautions are correctly stated to be those which are practicable or practically possible, taking into account all circumstances ruling at the time including humanitarian and military considerations. This wording reflects, verbatim, statement (b) made by the UK when it ratified AP I and the wording in Article 3(10) of the Amended Mines Protocol. It is a generally accepted interpretation of the word “feasible” in this kind of context. The need for immediate response when troops are in contact is, indeed, a good example of when military considerations may limit the precautions that are practically possible and is therefore usefully referred to in this part of the Manual text. The point is amply supported by the observations in footnote 106. In the next sentence the possibility that more information will be needed before approving planned future attacks is also suitably noted. It is also good that in the final sentence under this heading the potential need to take action “at the speed of relevance” is mentioned. This is an issue of growing practical important, as indicated earlier in this post. Doctrinal guidance, which may need to be classified, will be required so that the full range of required precautions, including those taken by personnel and those taken using technological systems, are set forth and understood by personnel who may be involved in countering threats at up to digital speed. It may be necessary to explain how “good faith” decision-making translates into automated or autonomous responses to, e.g., hypersonic, or cyber, digital speed threats. In my provisional view, the “good faith” obligation will apply to those who procure, test, develop, install, activate, de-activate, and monitor such systems and those who command such operations at various levels of command.

Doubt or Incomplete Information

Under the “Doubt or Incomplete Information” heading, the Manual correctly observes that the law of war allows for the possibility of doubt and of incomplete information. The first part of the next sentence is also correct in that the mere existence of some doubt as to whether the proposed target is a military objective does not necessarily preclude making the person or object concerned the target of an attack. In this respect, the materials cited in footnote 110 are authoritative. Indeed, one cannot help feeling that the Manual’s authors would have done well to elevate the Michael Schmitt and Michael Schauss quote to the main text and then immediately thereafter state the U.S. DoD understanding of what approach should be adopted when seeking to determine whether the amount of doubt constitutes too much doubt.

Instead, the rest of the sentence in the Manual consists of a proviso, i.e., that attacks where there is some doubt can proceed provided that decisions are made in good faith based on available information, that feasible precautions are taken to verify that the object is a military objective, and that action is taken consistent with the obligation to reduce dangers to civilians and civilian objects. This formulation might confuse the reader. The issue of doubt is relevant to the classification of a person or object as a military objective. Only if that classification is made by the commander, planner, or other decision-maker do the issues concerning discrimination, proportionality, and precautions arise. In the current formulation, the reader might be left with the mistaken impression, for example, that the obligation to minimise civilian casualties has something to do with the decision whether a person or object is a military objective and, thus, a lawful target for attack. Equally, the reader might be confused about the correct relationship between doubt, available information, and good faith.


Sub-paragraph under the title “Heightened Identification Requirements in Conducting Attacks” addresses the possibility that heightened standards of identification in excess of those required by the law of war might be applied by policy-makers to reduce the risk of incidental harm during attacks, referred to below as “precautions that are not legally required.” Implicitly, it is being suggested that the requirements mentioned later in the sub-paragraph are not required by the applicable law but would, instead, be applied as a matter of policy. The Factsheet cited in footnote 111, however, appears to refer to counter-terrorism operations which may not necessarily be conducted in the context of a state of war. One wonders, therefore, whether the reference in the footnote is necessarily strictly relevant.

The second sentence of the sub-paragraph gives an example of “precautions that are not legally required,” namely, “practices designed to mitigate cognitive biases that result in misidentification of targets in combat.” The associated footnote 112 gives examples of processes that can be employed to mitigate cognitive biases, including training, education and red-teaming. The reader will wonder how these processes and precautions relate to the obligation to take feasible precautions to verify that the person or object is a lawful target for attack. If the potential for cognitive bias is known, and if the potential for such bias to result in target mis-identification is known, and if the ability of certain processes such as training, education, and red-teaming to mitigate such bias is also known, surely the obligation to employ those processes in order to mitigate those undesirable outcomes is a part of the legal obligation to take feasible precautions. If that is right, the stated practices are not examples of precautions that are not legally required.

The final sentence of sub-paragraph is correct. Additional precautions that are undertaken as a matter of national policy do not affect the State’s obligation to fulfil the precautions that international law requires.


Sub-paragraph is entitled “AP I Provisions Regarding Doubt in Attacks.” It is sensible that the Manual draws attention to the obligations that bind the 174 States that are party to AP I. The first sentence of the sub-paragraph notes that in the context of attacks certain AP I rules provide for the application of civilian status in cases of doubt. That is, indeed, so. Article 52(3) of AP I, which states the AP I doubt rule with regard to objects, is then correctly quoted in the next sentence. The third sentence of the sub-paragraph then accurately quotes the doubt rule as to persons set forth in article 50(1) of AP I. Footnotes 114 and 115 provide useful additional information. The differing national interpretations of these AP I doubt rules are then noted, and the materials in the associated footnotes are helpful. As to the next sentence, the authorities cited in footnote 118 certainly suggest that the AP I doubt rules promulgated new rules of law when AP I was adopted. The issue that would interest the reader, of course, is whether in the view of the Manual’s authors, they represent customary international law now. The Manual states that they have been interpreted as being consistent with customary law and with the discussion in the relevant Section of the Manual. The “have been interpreted as” language might appear to some readers to be carefully avoiding making any clear statement as to whether the AP I doubt rules now reflect customary law, although there is certainly an inference to that effect. Clearly, only if the rules are customary do they bind States that are not party to AP I, such as the United States. The position seems to be made clearer in the next sentence where it is pointed out that the United States has accepted in a document referenced in footnote 120 the language in the CCW Amended Mines Protocol, which is identical language to Article 52(3) of AP I, and has interpreted that language as part of a prohibition that is “already a feature” of customary international law. It is, perhaps, on balance understandable that the departure in this section of Manual text from the previous assertion that the AP I doubt rules are not customary law should be expressed in somewhat cautious terms. It would, after all, be inappropriate boldly to state that the two rules are customary, given the action that may have been taken in the past in good faith by commanders and other decision-makers in accordance with the previous national position on the matter.

Likewise, the reader might legitimately wonder why the position on doubt with regard to objects has to be expressed in the Manual in a somewhat circuitous, potentially opaque, way. Why, in particular, could the discussion not simply start with a statement of the AP I Article 52(3) provision followed by a simple acknowledgment that the United States accepts this as a provision of customary international law. The sensible conclusion is that, as noted in the previous paragraph, there are sensitivities involved that preclude the making of such a clear statement. My view is that Article 52(3) does indeed reflect customary law.

Where the doubt rule as to persons is concerned, the next sentence in the Manual draws attention to previously expressed U.S. DoD concerns as to interpretations of Article 50(1) that would depart from customary law. The associated footnote, number 121, draws attention to a Chairman of the Joint Chiefs of Staff (CJCS) Memo in which the presumption of civilian status in AP I article 50(1) is rejected as it “provides an additional protection for guerrillas and other irregulars who may find it advantageous to be presumed a civilian rather than a combatant.” The U.S. DoD position on that matter is now made clear in the following sentence, where it is asserted that the discussion in sub-paragraph “reflects the DoD view of customary international law applicable to assessing whether persons or objects are military objectives, including in cases of doubt, when planning and conducting attacks.”

It is worth remembering that the referenced CJCS Memo dates from 1985. One wonders whether the increasing use of civilian contractors and, indeed, of private individuals to undertake activities that are linked to varying degrees with the conduct of hostilities reinforces or detracts from the U.S. concerns that were referred to in the Memo. In that regard, it must be borne in mind that for AP I States party, a civilian who take a direct part in hostilities loses his or her protection from being made the object of attack while doing so but remains a civilian.

Sub-paragraph 5.5.3

The final substantive amendment concerns the introduction of a new paragraph 5.5.3. Section 5.5 is entitled “Discrimination in Conducting Attacks” and paragraph 5.5.3 has the heading “Feasible Precautions to Verify Whether the Objects of Attack are Military Objectives.” The new paragraph starts with a clear statement that attack planners and decision-makers must take feasible precautions to verify. Footnote 175 usefully quotes the similar requirement set forth in AP I, Article 57(2)(a)(i). The next sentence rightly points out that taking these measures helps to implement the principle of distinction. The sentence thereafter then asserts, “Like other precautions in planning and conducting attacks, these measures are also part of the implementation of the principle of proportionality.” The potential problem with this sentence is that the measures being considered in this part of the Manual text are designed to verify whether the person or object is in fact a military objective. It is only once that determination has been made that the options for undertaking a planned attack can be considered, that the collateral damage to be expected from each of the attack options can be considered and weighed against the anticipated military advantage. It would, perhaps, be more accurate, therefore, for the sentence to read, “Once the determination that the person or object is a military objective has been made, these measures may also provide information that can help to determine whether a planned attack will satisfy the proportionality rule.” In the next sentence it is indeed right to note that these measures help to reduce uncertainty and reinforce military effectiveness in the manner suggested.

The precautions that are feasible to take will, indeed, depend greatly on the circumstances and footnote 181 suitably supports that observation. In the ensuing few sentences the point is usefully developed, noting the significant difference in this regard between troops in contact situations and deliberate targeting. It would, perhaps, have been useful to add into the main text just before footnote 184, “The verification measures that are in fact adopted should be the best means that are available under the prevailing circumstances, taking into account the immediacy of any potential threat.” The second sentence of the extract from the German Manual cited in footnote 183 neatly encapsulates the relevant point and would, it is suggested, enhance the reader’s understanding. It would also provide a link to the valid point made in the next sentence, namely that the law of war does not indeed preclude Commanders and others from taking decisions at the “speed of relevance,” and this author reminds the reader of the point made earlier concerning the rapid threats made possible by modern technologies.

The remainder of this sub-paragraph gives a non-exclusive list of possible kinds of precaution that may, depending on the circumstances, be feasible and, thus, required. The reader must, of course, appreciate that the precautions in the list are no more than examples, but the introductory words do make that sufficiently clear. The references in footnotes 187 and 188 are useful and relevant.

Concluding Observations

The remainder of the changes introduced in 2023 concern the integration of cross-references, organisational refinements, and the correction of minor errors. These will not be considered in this discussion.

In bringing this piece to an end, it is appropriate to reflect briefly on the observations that certain other commentators have already made on the 2023 amendments. Michael Meier has pointed out that the DoD Manual’s position on the presumptions of civilian status had been criticised by commentators. Michael Meier welcomes the Manual’s altered position on the presumptions of civilian status as it more accurately reflects U.S. practice in the treatment of civilians and civilian objects, because “it clarifies that section reflects the DoD view of customary international law applicable in cases of doubt when planning and conducting attacks,” because it harmonises the U.S. position with that of allies and partners, and because it “corrects the mis-statement of international law that no legal presumption for civilian status exists for persons or objects.”

Professor Michael Schmitt has looked in detail at the doubt issue in numerous writings, most recently this month. He identifies as a key question “when is proceeding in the face of uncertainty unreasonable and, therefore, unlawful?” After mentioning a number of situational examples in which such uncertainty may arise, Professor Schmitt concludes, correctly in my view, “The question is not whether a one-size fits all threshold has been reached, but instead, whether the attacker is confident enough in the circumstances to mount the attack lawfully.” He concludes that the changes made to the DoD Law of War Manual are positive, and I share that view.

I do, however, hope that some at least of the comments and suggestions I have made with regard to the drafting of the amended Manual provisions are found to be useful. I nevertheless commend the U.S. Department of Defense for its on-going effort to keep its Manual up to date.


Air Commodore William H. Boothby retired as Deputy Director of Royal Air Force Legal Services in July 2011. He is Honorary Professor at the Australian National University and also teaches at the University of Southern Denmark and at the Geneva Centre for Security Policy.


Photo credit: Tech. Sgt. Daniel Asselta