Introducing the Canadian Defence and Security Network

by Stephen M. Saideman

The time has come to roll out the Canadian Defence and Security Network.   We have been working on funding the CDSN for several years, so we are elated and just a wee bit anxious. We have built an excellent team of scholars and defence scientists to lead the effort and already have a terrific staff to do much of the heavy lifting and day-to-day management. In addition to that, we have so many partners in a variety of sectors in Canada and beyond who strengthened our application through the commitments they have made.   I am so very grateful for the work done thus far and the work to be done by our leadership team (David Bercuson, JC Boucher, Andrea Charron, Irina Goldenberg, Phil Lagassé, Anessa Kimball, Alex Moens, Alan Okros, Stéphane Roussel, Stéfanie Von Hlatky, and Srdjan Vucetic),* the folks at CDSN HQ (Jeffrey Rice, Melissa Jennings, Alvine Nintai), the people at NPSIA, our dean (Andre Plourde), grants facilitator extraordinaire (Kyla Reid), other folks at Carleton including our VP for Research (Rafik Goubran), and our partners and participants.  I look forward to working with these terrific people along with others who join our efforts.

* Note we have plenty of Francophones on our leadership team that will help compensate for my being linguistically lame.  While my blog here is unilingual, we will try to make sure that much of our stuff will be accessible in both official languages.

Of course, as you are reading this, you are asking yourself: what is the Canadian Defence and Security Network and what is it supposed to do?

It is a partnership involving academics at both civilian and military universities, units within the Canadian Armed Forces, elements of the Department of National Defence, think tanks, advocacy organizations, a survey firm, and more.  We have a set of common objectives:

  • To create a coherent, world-class research network.
  • To advance our understanding of defence and security issues.
  • To tailor research initiatives to provided evidence-based knowledge to inform policy-making
  • To facilitate the transfer of knowledge and data across various divides.
  • To improve the defence and security literacy of Canadians (and beyond).
  • To build the next generation of experts with an emphasis on equity, diversity and inclusion.

How will we reach those objectives?  The CDSN will focus on five themes to coordinate research efforts–military personnel, defence procurement, operations, civil-military relations, and security–while also providing resources via our headquarters to assist its members and its partners to collaborate and amplify their work.

To provide an example, one can imagine an event organized by scholars in Kingston or Calgary.  The CDSN Headquarters (based at CSIDS at NPSIA) will help provide contacts to reach out beyond the networks of the organizers, assist if grant-writing is required, will help publicize the event through the CDSN’s social media efforts (yes, we have some experience in that stuff) including a blog, twitter account, and podcast, and then after the event, provide a repository for the data generated, the papers and policy briefs that are produced and spread the findings via our website.

Please note, despite our years of prep work, we are very much a work in progress.  We are officially launching the CDSN on May 24th, and our first major event will be the Kingston Conference on International Security (KCIS) in early June.  While that event has been a great conference involving not just Queens’s Centre for International and Defence Policy but also the NATO Defence College and the US Army War College’s Strategic Studies Institute and various CAF elements, we hope that the CDSN will help KCIS have a broader reach across Canada, and it will probably provide our first podcast content!  We will also be supporting the Women in International Security-Canada Annual Workshop later in June.

For our first year, we will be focusing mostly on developing our infrastructure and figuring out how to help the various members of the CDSN community.  In years 2-7, we will have thematic workshops on our five areas of research; book workshops for junior scholars; post-docs; surveys of the Canadian public; network analyses; summer training institutes for scholars, military officers and policy officers; an annual conference; defence fellowships for military officers; and capstone events that will bring the best young presenters from events across Canada together to present to defence policy-makers.

Our twitter account is: https://twitter.com/CdsnRcds.  The website will be populated as time goes on, and we will certainly have facebook, instragram and other social media accounts that we will be announcing over the next few months.  Our logos are a work in progress, but this is what we have thus far:

If you are interested in joining our efforts or have questions, shoot us an email at cdsn-rcds@outlook.com.

 

 

Pardons of War Criminals and the Future of Allied Cooperation

by Stephen M. Saideman

The weekend’s news suggests that President Donald Trump’s pardon of a war criminal, former Army officer Michael Behenna, is not going to be a one-off thing but part of a broader trend of pardoning those accused or convicted of war crimes.  While this policy presents challenges to the American armed forces—endangering discipline and cohesion, my focus here is on the impact on present and future allies.  Simply put, this new stance will make it much harder for many countries to join the US in any future military campaign.  Here, I take a quick look at what the allies provide and then focus on how pardoning war criminals is likely to affect future military cooperation.

Continue reading

Japanese Aircraft Carrier’s: Asking Questions

by Stephen M. Saideman

From the National Interest.

The news of the month, Asia-Pacific-wise, is that Japan has admitted that its escort destroyers (and other names for its helicopter carriers) are going to be converted into aircraft carriers via some modifications and with purchasing the version of the F-35 that can take off from a short deck.  Why?

That is, why is Japan doing this?  I raise that simple question.  Folks think the answers are obvious:

a) China is bad and spending heaps on a larger navy

b) Trump is unreliable so Japan needs carriers more now than before.

And my response to these answers is this: how do carriers solve these problems?  Maybe they do, maybe they don’t.  But my study of Japanese civil-military relations suggests that damn few civilians are asking any questions at all.

Continue reading

Bullies in International Relations

by Stephen M. Saideman

The past few years challenge much of the conventional understanding of international relations.  One of the big lessons from the IR scholarship of the 1970s is that the nature of international relations is that threats and bullying don’t work.  As Robert Jervis discussed it, the world can be either a constant chicken game or a repeated prisoner’s dilemma–aka deterrence vs. spiral model.  In short, is international relations an environment (a system!) where countries cave into threats or do they balance against them, that those who believe that pushing countries around are usually confronting with coalitions created by such bullying.  Kaiser Wilhelm, as IR scholars use as a example, threatened everyone, hoping that they would back down.  Instead, these countries solidified their alliances and showed up in Europe in August 1914.  Oops.

Over the past several years, we have seen a series of countries engage in bullying behavior–Russia, Saudi Arabia, Trump’s US and China.  Russia has wielded nuclear threats to encourage Europeans to not deploy troops to the Baltics and to dissuade them from supporting Ukraine.  How has that worked so far?  Saudi Arabia has seemingly become unhinged as of late, overreacting to Canadian discussions of Saudi human rights and all but warring upon Qatar.  Trump, well, is a bully, so we ought not be surprised by his threats nor by his ignorance of IR scholarship. Threatening the allies has led them to ponder hedging and alternatives.  He might think the North Koreans have submitted after last year’s threats, but I am pretty sure the North Koreans think they have the upper hand.

The big surprise, to me anyway, is China.  China has manged its rise so very well in large part because it has mostly wielded a velvet fist.  Yes, it has buzzed American planes and ships, had friction with Indonesia, and other stuff.  But generally, the China of the 2000s and early 2010s has been replaced by a more aggressive and obnoxious China.  The tiff with Canada is importance since Canada was the western democracy least likely to object to the Huawei company getting inside Canada’s 5G.  Well, not any more.  The current standoff is causing Canadian parties to rally against China–who is arguing now that Canada should submit?  Moreover, a conversation with a European diplomat today reminded me that Canada has more influence than folks think.  Not necessarily to push China back into the straight and narrow but to serve as bellwether.  If  a country has a problem with the US or EU, well, those are powerful entities that can antagonize.  But a country has a problem with Canada?  That suggests that the particular country is problematic… and, jeez, is China problematic these days.

I am not a China expert so I don’t really know what is driving China to behave this way.  I would guess domestic politics and nationalism (populism?  Not quite).  But everything I have learned in my career tells me that China’s choices now are self-destructive–that being aggressive does not pay in the long run.  That bullying is counter-productive.  Perhaps China is encouraged because the US led by Trump is so incompetent and unreliable, which means balancing will be late, inept and weak.  But it is still a dumb move–the Chinese have been gaining strength with little opposition because they were not overly aggressive.

The thing about IR theory stuff–it didn’t say that bullying didn’t happen. It just said it was not productive.  So, the question for future IR scholars, if we live so long, is whether China, Russia, Saudi Arabia and Trump’s US are punished or not.  We shall see.

 

NATO and the International Order: It was Kind of Nice, Wasn’t It?

by Stephen M. Saideman

People have been asking me lately–what is the big deal with this international liberal order?  What has it ever done right? What has it ever given me? There are lots of pieces to it, but I am focused on NATO for obvious reasons, including my assignment at next week’s Kingston International Conference on Security.

So, here’s Mattis’s quote from the NAC (North Atlantic Council) Defense Ministerial:

and my reaction.

I used to scoff at the usual NATO existential crisis stuff–that NATO needed a reason to exist in the aftermath of the Soviet collapse, that there was some conflict within that might lead to the alliance breaking up, etc.  But now I am in the club of those who fear for NATO’s future.  Why?  Trump.  It is that simple.  Putin actually did more for NATO unity in 2014 than anything else by making folks remember NATO’s day job–keeping Europe peaceful and, as a result, prosperous.  But his gambles on Trump, on Brexit, on supporting right wing aspiring autocrats (Orban of Hungary, Erdogan of Turkey, etc) have worked out.

The alliance has worked and changed our conception of alliance not just because it is far more institutionalized than any other alliance past or present, but because all of it relied on largely shared values.  Not just democracy but democracy with embedded liberalism–that governments played a role in adjusting to international shocks, made easier by international cooperation.

And now is a splintered G-7 meeting due to Trump using “national security threat” to play a particularly problematic card–to impose tariffs on allies without the consent of Congress.  To be clear, this is the opt out card built into the agreements.  He does not really believe that these countries or their exports to the US are any kind of threat, but he does not believe in norms, rules or the future. So, Trump has used this exception, antagonizing everyone except maybe the Italians (their own populist election results are handy for self-destruction).

So endeth the shared values.  Orban has already promoted illiberal democracy, and Trump would too if he could articulate anything (note that Gorka is back, and Gorka is a living embodiment of Orban’s illiberal democracy).  True, Trump is not the US, but he is, alas, 40% of it, and the GOP seems ok with selling out American values for tax cuts and court seats.  So, even if/when the Democrats come into power, they will not be able to reassure the Europeans and the Canadians.  After all, this big split is the most significant … since the last Republican president and the misconceived Iraq war of 2003.

So, how can NATO provide security by reassuring nervous members and deterring adversaries?  The lack of common values undermines NATO credibility–will the US show up if Russia does something?  Perhaps not since Trump is now trying to get Russia back into the G8 despite everything Russia has done since seizing Crimea.

NATO isn’t dead, and I hope to see signs of life when I go to the expert side party at the summit next month. But NATO is far from healthy, and I worry that we soon look back at those 70 years Mattis speaks of and wistfully remember the good old days.  Maybe the good old days weren’t as good as they seem, as Billy Joel reminds us, but they were better than the days before that–WWI, WWII and all that.

 

Much Learning in Two Weeks in Korea

by Stephen M. Saideman

Fun times in Korea, eh?  I was really struck during my two weeks there of a split in opinions–most of the folks I met were “cautiously optimistic” about the situation, that the Trump-Kim summit might lead to a significant improvement in regional tensions, while other folks were in the “ruthlessly pessimistic” camp.  And I was a member of the latter.  Why?  Because TeamRP just could not see anyway for North Korea to “denuke” in any meaningful way when the US had, ooops, done some regime change on Libya.

So, I get back to North America and notice that Bolton has been talking about the “Libya Option” seriously, which did ultimately send the desired signal (if Bolton does not want peace) to the North Koreans.  So, the North Koreans have said that they had no intention of trading their nukes for economic assistance.  That, along with the earlier announcement that that they were skipping a meeting since the US and ROK were not cancelling a key military exercise, reminded us that North Koreans have always been the most obnoxious trolls in International Relations (sorry, John and Stephen).

So, folks are having an epiphany–negotiating with North Korea is hard, and they aren’t giving up their nukes.  I had a bit of a different Korean epiphany thanks to some sharp outsiders (Canadians and Americans who took me out for drinks and bbq:

American troops have long been based in South Korea to do two things: deter the North Koreans and reassure the South Koreans.  Standard tripwire type stuff.  Now, things have flipped as smart South Koreans want the Americans to stay to deter an American attack on North Korea.  Yeah, that seems backwards, but the idea is that Trump would not attack North Korea with so many Americans in harm’s way (is that wishful thinking rearing its ugly head again?).  That Trump would have a freer hand if the Americans were no longer down range of North Korean artillery….

Before I left for South Korea, I thought that the likely outcomes from a KJU-Trump summit would be in decreasing likelihood:

  1. A modest agreement, such as NK agrees not to test any more nukes (its test area is broken and other new nuclear powers tested six nukes, so a convenient time to give away this chip) and US promises to de-escalate a smidge.  Trump would come home, declaring he solved the Korean problem, and the pundits/press would buy it, but not much would have really changed.  Woot!
  2. NK agrees to give up its ICBM capability, Trump agrees to reduce or even eliminate US forces in South Korea, so NK gets not only recognition of being a nuclear power but decoupling of South Korea and Japan from US.
  3. Trump and KJU yell at each other as each is upset that they don’t have a common understanding of “denuclearization.”  So, the road to war is a bit clearer, and John Bolton does a happy dance.

Now?

  1. No meeting as NK does not want to signal that it gave in to “massive pressure” from US.
  2. No meeting as Trump realizes he can’t get the Nobel Prize.
  3. A meeting with much reduced expectations–perhaps freezes of NK’s weapons in exchange for US promising not to regime change (which is believed by none).
  4. War.
  5.  A meeting, then war.

So, yeah, not great.  Are things clearer now than two weeks ago?  Not sure.  I do think Team Relentless Pessimism is feeling pretty good about feeling pretty bad. Woot?

 

What Do We Know About the Kurdish Referendum

by Stephen M. Saideman

Not much as I haven’t studied the Kurds.  McGill Phd David Romano has studied them a great deal, and there are others who are far more expert.  However, I do know something about separatism, referenda and irredentism, so here’s what I think:

  1. Separatism is not as contagious as advertised. The only folks who really get encouraged by an effort, successful or otherwise, are those who are kin.  Everyone else is far more focused on their own incentives and constraints.  They will learn from the examples elsewhere whatever lessons that support their pre-existing inclinations.  Yes, I was a fan of confirmation bias long before it was cool.
  2. The Sunnis will not be pleased.  It is hard enough for two smaller groups to attempt to balance the Shia in whatever semi-democratic institutions, but with Kurds leaving, Sunnis are dwarfed by Shia.  Hard to craft a democracy or anything else that gives Sunnis some chance of not being dominated.  So, yeah, Kurds leaving would screw Sunnis just as Slovenia screwed the Bosnians.  Everything old is new again.
  3. Irredentism is not in the cards.  Sure, one could talk about a Greater Kurdistan, but which Kurds get to rule it?  Milton (and Khan) was right: better to reign in hell than serve in heaven.  So, no, despite what Turkey might say, there will be no significant effort at Greater Kurdistan.
  4. I am not a fan of secession from advanced democracies–the costs of changing are too high and always underplayed by the secessionists, democracy depends on losers staying in the system, and usually there are ways to get what you need, if not what you want.  But the Kurds have some reasonable grievances, starting with how they can’t trust the Shia dominated government of Iraq.
  5. The timing makes sense–Kurdish strength is at an alltime high given that the US, Canada and others have armed and trained the Kurds.  Those efforts are already declining now that ISIS has been mostly removed from the Kurds’ neighborhood.  Iraq is still weak due to the ongoing war with ISIS, so now makes sense….
  6. But a referendum does not mean independence.  It can mean a process, a bargaining process that can take quite some time.  The question of violence really now depends on what the Iraqi government will do.  Governments generally don’t let secessionist regions leave–lots of work on this especially by Monica Toft.
  7. Countries will support whichever side they have ethnic ties (article version).  If no ties, then other interests, such as seeking stability will kick in.  The one thing, for damn sure, is that countries will not be deterred by their own vulnerability to separatism.
  8. Turkey will overreact.  Duh.

What does it mean for the war against ISIS?  Damned if I know.  Anyhow, my past work suggests this will be both better and worse than what the pundits say.  Woot?

Nuclear Logics: Uncertainty Is Not Necessarily Bad

by Stephen M. Saideman

Reading Andrew Coyne’s piece on North Korea and the need for Canada to join the American Ballistic Missile Defence System reminded me of lots of old deterrence theory stuff.  The piece raises good questions about the reliability of the key actors, especially Trump, but confuses what is necessary for deterrence.  Still, there are some problems that we need to think about.

The big problem in the piece is that Coyne thinks that the American commitment to defend its East Asia allies is now largely unbelievable with the North Koreans developing the ability to strike the continental US (and Canada).  It is true that the US, under several Presidents, has failed to deter the North Korean effort to develop both nuclear weapons and the missiles to deliver them.  But deterring their effort to develop some deterrence and deterring an attack on allies are two different things.  Coyne is right to point out that extended deterrence (don’t attack my allies or else) is less believable than regular, vanilla deterrence (don’t attack me or else).  The threat to start or expand a nuclear war is problematic in either case, but seems a bit more believable if it is in retaliation for a big attack on the homeland.

The key is that for deterrence to work, the side being deterred (North Korea in this case for the moment) does not need to be certain that a counter-strike would happen.  They just need to think that there is some possibility of such a response.  Why?  Because the costs of nuclear war are so very high, if one does the probability math (probability of x times value of x), the prospective costs of attacking first are simply too high compared to the status quo (.01 times infinity = infinity) …. as long as the status quo is bearable (which is why we have to stop threatening regime change).  We do not have to convince North Korea that a retaliatory strike is certain if North Korea attacks South Korea and/or Japan, but that it could happen.  The placement of US troops in South Korea is far more about being a tripwire to raise the probability of the US responding than actually defending South Korea in a conventional attack.

Again, one might say that this is not sufficient, but the key to nuclear threats is that classic Schelling phrase: a threat that leaves something to chance.  One does not have to threaten, for instance, total nuclear annihilation of North Korea crosses the Demilitarized Zone–one just has to threat to start a process that might lead to things getting out of hand and ultimately leading to nuclear war.  This was the old extended deterrence logic for Europe and Asia during the cold war.

Certainty?  That is for allies.  That is, the tripwires and all the rest over the years are mostly aimed at reassuring allies.  The enemy is deterred by a modest chance of the US responding, of sacrificing Chicago for Bonn or now Seattle for Seoul.  The allies?  They are very nervous and require a great deal of assurance.  Ballistic missile defense both in the region and in the US might assure them somewhat–that the US can stick around and meet its commitments knowing that it is protected.

Except for one thing: BMD may be at best a coin flip.  We have lots of uncertainty about whether the efforts to invest in destroying missiles in flight have produced anything remotely reliable.  Again, that is ok from a deterrence perspective–uncertainty is not bad.

While I think that joining the US BMD program makes sense, my reasons do not center on the NK nuclear threat.  The US would try to stop an attack on Vancouver or Toronto since they are very close to American cities whether Canada is in or out of the BMD program.  And North Korea is not going to be gunning for Calgary or Edmonton.  North Korea barely notices Canada, and, given its small supply of nuclear arms, it will not be aiming at Canada.  The BMD arguments I buy have more to do with building a robust NORAD that addresses a variety of threats in the 21st century, and strengthening a key US-Canadian institution in these uncertain times.

While we should all doubt Donald Trump, I am far more worried about his starting a process that leaves something to chance via a small strike at North Korea’s missiles or at its leadership than I am about his not responding to a North Korean attack.  Yes, we are now deterred from attacking North Korea, but that has been true since it developed the capability level Seoul with its artillery.  Yes, we have much to worry about, but then so does Kim Jong Un.  If he wants to survive, he will avoid a process that might lead to escalation.  The costs of being wrong are just too high.

Old Deterrence Logics Still Apply

by Stephen M. Saideman

Reading Andrew Coyne’s piece on North Korea and the need for Canada to join the American Ballistic Missile Defence System reminded me of lots of old deterrence theory stuff.  The piece raises good questions about the reliability of the key actors, especially Trump, but confuses what is necessary for deterrence.  Still, there are some problems that we need to think about.

The big problem in the piece is that Coyne thinks that the American commitment to defend its East Asia allies is now largely unbelievable with the North Koreans developing the ability to strike the continental US (and Canada).  It is true that the US, under several Presidents, has failed to deter the North Korean effort to develop both nuclear weapons and the missiles to deliver them.  But deterring their effort to develop some deterrence and deterring an attack on allies are two different things.  Coyne is right to point out that extended deterrence (don’t attack my allies or else) is less believable than regular, vanilla deterrence (don’t attack me or else).  The threat to start or expand a nuclear war is problematic in either case, but seems a bit more believable if it is in retaliation for a big attack on the homeland.

The key is that for deterrence to work, the side being deterred (North Korea in this case for the moment) does not need to be certain that a counter-strike would happen.  They just need to think that there is some possibility of such a response.  Why?  Because the costs of nuclear war are so very high, if one does the probability math (probability of x times value of x), the prospective costs of attacking first are simply too high compared to the status quo (.01 times infinity = infinity) …. as long as the status quo is bearable (which is why we have to stop threatening regime change).  We do not have to convince North Korea that a retaliatory strike is certain if North Korea attacks South Korea and/or Japan, but that it could happen.  The placement of US troops in South Korea is far more about being a tripwire to raise the probability of the US responding than actually defending South Korea in a conventional attack.

Again, one might say that this is not sufficient, but the key to nuclear threats is that classic Schelling phrase: a threat that leaves something to chance.  One does not have to threaten, for instance, total nuclear annihilation of North Korea crosses the Demilitarized Zone–one just has to threat to start a process that might lead to things getting out of hand and ultimately leading to nuclear war.  This was the old extended deterrence logic for Europe and Asia during the cold war.

Certainty?  That is for allies.  That is, the tripwires and all the rest over the years are mostly aimed at reassuring allies.  The enemy is deterred by a modest chance of the US responding, of sacrificing Chicago for Bonn or now Seattle for Seoul.  The allies?  They are very nervous and require a great deal of assurance.  Ballistic missile defense both in the region and in the US might assure them somewhat–that the US can stick around and meet its commitments knowing that it is protected.

Except for one thing: BMD may be at best a coin flip.  We have lots of uncertainty about whether the efforts to invest in destroying missiles in flight have produced anything remotely reliable.  Again, that is ok from a deterrence perspective–uncertainty is not bad.

While I think that joining the US BMD program makes sense, my reasons do not center on the NK nuclear threat.  The US would try to stop an attack on Vancouver or Toronto since they are very close to American cities whether Canada is in or out of the BMD program.  And North Korea is not going to be gunning for Calgary or Edmonton.  North Korea barely notices Canada, and, given its small supply of nuclear arms, it will not be aiming at Canada.  The BMD arguments I buy have more to do with building a robust NORAD that addresses a variety of threats in the 21st century, and strengthening a key US-Canadian institution in these uncertain times.

While we should all doubt Donald Trump, I am far more worried about his starting a process that leaves something to chance via a small strike at North Korea’s missiles or at its leadership than I am about his not responding to a North Korean attack.  Yes, we are now deterred from attacking North Korea, but that has been true since it developed the capability level Seoul with its artillery.  Yes, we have much to worry about, but then so does Kim Jong Un.  If he wants to survive, he will avoid a process that might lead to escalation.  The costs of being wrong are just too high.

Crisis in US Civil-Military Relations? Not Yet

by Stephen M. Saideman

Yep, no process, no policy, no implementation.  I wrote yesterday that Trump’s transgender in the military “policy” would depend on how the military would feel about implementation.  Well, from the very top, the attitude is: wait and see.  More than that: a smidge of contempt seems to be in the reaction:

Dunford has informed service
members that there will be “no modifications to the current policy until
the President’s direction has been received by the Secretary of Defense
and the Secretary has issued implementation guidelines.”

“In the meantime, we will continue to treat all of our personnel with
respect,
” Dunford wrote in a memo to the military that was obtained by
CNN. “As importantly, given the current fight and the challenges we
face, we will all remain focused on accomplishing our assigned
missions.”CNN (I would have cited NYT but they don’t let me cut and paste!)

If Dunford were General (ret.) Kelly of Homeland Security, he might have taken the tweet and ran with it, as Kelly enforced an immigrant ban with very little backing it up.  Dunford, like the other active senior officers, has opposed kicking transgender people out even as they hem and haw on how to deal with recruiting.  So, this agent has preferences that are distinct from the principal and, as a result, does not imagine what the tweet actually means, but instead asks for the paperwork to be done.

And, yes, DC runs on paperwork …. or Word docs shipped around town as attachments to emails (yes, on the classified servers mostly).  Since Mattis has thus far been silent (did he say anything while I was at Costco?), Dunford went ahead and interpreted how far he could go and went pretty far.  I had some responses on twitter asking for him to do more.  Such folks don’t understand civil-military relations–that civilian control of the military means that the civilians have the right to be wrong (which they are here), that the military must obey clear orders.  But they can fudge implementation if the orders are not clear or are not handed down through the chain of command.  Dunford could have started a process to weed out the transgender soldiers, sailors, marines and aviators, but chose not to do so.  This is kind of a work-to-rule thing, where resistance of this form is merely following the rules.  Trump would need to find another general who is more enthused about discrimination to get faster action.  Firing a Chairman for this?  Unlikely.

Finally, it is good to see someone indicate that a tweet may be a policy direction but is not a policy itself.