This week there was another inconclusive round of negotiations between London and Brussels on the Northern Ireland Protocol, which remains the biggest source of friction between the EU and UK.
At the heart of the friction is a claim by London that the EU is overly zealous in applying the Protocol, and a counterclaim that the UK knew exactly what it signed up to.
It is part of a treaty (the Withdrawal Agreement) and so judging who is right and who is wrong is not always easy if you have to reach for a dense legal text.
And yet clauses such as "Article 16" have become part of the day-to-day political jousting.
As it happens, Oxford University Press (OUP) has just published a major guide to how the Protocol should be read and understood.
The UK-EU Withdrawal Agreement: A Commentary is a line-by-line explainer of each and every part of the Withdrawal Agreement, with a target readership of academics, law practitioners, governments and courts.
The fact, however, that it was written by EU and UK officials who actually took part in the negotiations, including those who co-drafted the Protocol, gives the work a genuine resonance.
They include Catherine Adams and Daniel Denman, both directors in the UK government's Legal Department and Thomas Lieflӓnder, a senior official in the European Commission and member of Michel Barnier’s Brexit Task Force.
Of the nine authors involved only one, Manuel Kellerbauer, a specialist in EU law, was not involved in the negotiations.
However, there is an important disclaimer at the outset: "Nothing in this commentary should be construed as representing the positions of either the UK government or the EU institutions".
According to one participant, the commentary is "neither the EU’s nor the UK's interpretation" but a technical analysis by lawyers who happen to have been involved and who "know the text very well".
The section covering the Northern Ireland Protocol takes up 80 of the 543 pages.
It is hard going, but there are a number of immediate impressions.
The commentary asserts that this was an extremely complex and unprecedented arrangement, a messy solution to the impact of Brexit on the fragile arrangements on the island of Ireland.
Once the UK (reluctantly) agreed that keeping Northern Ireland in the single market was the least worst option, the EU side made sure that the arrangements were legally stern, robust and enduring.
From the EU's point of view, it was outsourcing control of its external border to a third country (the UK) so the arrangements had to be precise.
If the UK decided to trigger Article 16, as it has relentlessly threatened, it would most likely be to suspend those parts of the Protocol which essentially create the Irish Sea border
The commentary conveys the view that each Article was carefully weighted and balanced. There is frequent "read-across" to the EU’s founding treaties.
When the text is examined in detail it is no surprise that the EU has resisted any attempt at taking bits and pieces out here and there - as the UK Command Paper of July 2021 demands.
The EU would prefer to depict the current negotiations as talks about finding pragmatic solutions, but within the framework of the Protocol.
That said, the OUP commentary frequently acknowledges the sensitivities on the UK side.
Part of the necessary political dexterity at the time involved camouflaging - or at least not putting in bright neon lights - the fact that Northern Ireland would be de facto staying in the EU’s single market.
As such, the Protocol is littered with signposts, referring to one or other article from EU treaties and which will send the reader off to annexes containing more detail about how EU law will still apply in Northern Ireland.
For that reason, according to the commentary, both sides decided to front-load the Protocol with strong reassurances (which run through Article 1) that it did not encroach upon the territorial or constitutional integrity of either Northern Ireland or the UK.
Nor did the Protocol have to wait for the outcome of the future trade agreement. It was essentially intact.
"Implicit in this language," writes Thomas Lieflӓnder, the senior Commission official involved in the negotiations, "is the appreciation of the Parties that…the arrangements…are both necessary and sufficient to achieve its objectives.
"[They underline] that from the perspective of the EU–UK relationship, the Protocol provides a complete response to the unique situation on the island of Ireland and is not dependent on any future agreements."
This is at odds with the UK government’s current position, which is that parts of the Protocol have been superseded by the Trade and Cooperation Agreement (TCA) concluded in December 2020.
The UK Command Paper argues that Northern Ireland remaining part of the EU’s state aid regime was the "result of the fact that the Protocol was negotiated in 2019 when neither the UK nor the EU knew the nature of commitments to be made on subsidy control in the future Trade and Cooperation Agreement or indeed whether it would be agreed at all."
Article 16 is the best known article, of course.
The commentary breaks down how the article came about and what it is for - and the way much of it is open to interpretation.
The Article is closely modelled on a clause in the EEA treaty (the European Economic Area brings together the EU and the European Free Trade Association (EFTA)) which provides safeguards if a signatory country is hit with imports which unexpectedly damage a sensitive sector.
Article 16 allows either side to take temporary and limited safeguard measures, while the other party can then take "proportionate rebalancing measures."
However, the problem with Article 16 is that it is somewhat vaguer than the EEA safeguards. Those safeguards spell out that damage has to hit a particular sector or region, but Article 16 does not.
That means Article 16 has a wider scope. Furthermore, unlike the EEA approach, Article 16 adds "diversion of trade" to the list of circumstances which might allow a trigger.
But there is strong language suggesting it is not to be triggered lightly.
Triggering Article 16 is constrained by a combination of qualifiers - "serious [...] difficulties...likely to persist…strictly necessary" - and conditions that both sides must follow should the trigger be pulled: immediate consultations, ongoing consultations and review of the decision.
When Boris Johnson took over, he ditched Theresa May's idea of a UK-wide customs union with the EU, which would have dispensed with customs formalities on the Irish Sea
The Article, indeed, makes it clear that any move to trigger should disturb the balance of the Protocol as little as possible, and that it cannot be used "if the difficulty in question is a logical and unavoidable consequence of the application of the Protocol."
The way it was drafted, writes Lieflӓnder, underlines the fact that Article 16 is "truly a measure of last resort".
Despite this, Article 16 is strangely vague on what constitutes "serious economic, societal or environmental difficulties", or what would qualify as a sufficient "diversion of trade’".
Of course, if the UK decided to trigger Article 16, as it has relentlessly threatened, it would most likely be to suspend those parts of the Protocol which essentially create the Irish Sea border.
These are found in Articles 5-10, and they take up a large part of the OUP commentary.
These Articles deal with the reality that goods leaving GB and entering Northern Ireland would be leaving one internal market and customs regime (the UK) and entering another (the EU single market and customs union).
This is where the friction happens, the checks and controls which have led to a sustained political crisis in Northern Ireland, and a sour standoff between the EU and UK.
The commentary provides a poignant reminder that the friction need not have been so hard on the Irish Sea.
While Lieflӓnder says that much of Article 5 can be traced directly back to the Joint Report of December 2017, where the UK accepted the "full alignment" of Northern Ireland with the EU single market, much of the negotiation around Article 5 happened late in the day.
This is because when Boris Johnson took over, he ditched Theresa May’s idea of a UK-wide customs union with the EU, which would have dispensed with customs formalities on the Irish Sea.
Therefore Article 5 "has to be seen as an attempt to grapple with some of the consequences of moving from the UK-wide customs arrangements… to the Northern Ireland-only arrangements enshrined in the final version".
Article 5 clarifies that even if there were no tariffs due on goods moving from Britain to Northern Ireland, customs formalities would still apply.
The Protocol left it to the EU-UK Joint Committee to work out how exactly you could prove that goods weren't at risk of crossing the border
It also sets out the whole notion of "at risk", which bedevilled a lot of the negotiations over how the Protocol should be implemented, before it took effect on 1 January 2021.
If goods were staying in Northern Ireland then they should not attract tariffs (this was ahead of the TCA which established a tariff-free trade deal). If goods were "at risk" of crossing the border, then they would attract tariffs.
However, the way Article 5 was drafted, the default was that goods were guilty of being at risk of crossing the border until proven innocent.
"By default," explains Lieflӓnder, "all goods entering Northern Ireland have to be treated as being at risk… this presumption has to be actively rebutted by establishing that certain conditions negating that risk are met."
This became a very sore point between London and Brussels in 2020. However, this legal commentary makes it clear that that was the intention.
The Protocol left it to the EU-UK Joint Committee to work out how exactly you could prove that goods weren’t at risk of crossing the border.
After a tortuous set of negotiations in the autumn of 2020, the Joint Committee, under the UK’s Michael Gove and the EU’s Maros Sefcovic, agreed a formula.
Goods coming directly from GB which were solely destined for Northern Ireland supermarkets and then consumers, would be deemed largely not at risk.
A trusted trader scheme, backed up by labelling and the heavy exchange of data, would provide further reassurance that goods weren’t leaking into the single market.
Even then, the EU was concerned that this was too "light touch". As a result, both Gove and Sefcovic agreed a sunset clause that would scrap the trusted trader scheme by 1 August 2024, if either party was concerned about a "significant diversion of trade, or fraud or other illegal activity."
As of this week, while there has been a lot of technical work on customs, there is still no breakthrough
This, says Lieflӓnder, was an emergency brake for the EU in case the Gove-Sefcovic deal turned out "not to be effective at protecting the integrity of the EU single market and customs union".
In the negotiations currently underway between Sefcovic and Gove’s (second) successor Liz Truss, the UK wants to go well beyond the December 2020 deal.
Essentially, Britain wants there to be no checks at all on goods that are clearly going to stay in Northern Ireland, with compliance and traceability being outsourced to traders themselves.
The EU has shifted its position. In its October 2021 proposals it signalled that it could live with a greater level of risk (although not much).
The Commission paper on customs talked of "genuine simplifications" that could cut formalities by 50%, taking into account the nature and size of the goods, whether those affected were traders, consumers, retailers, wholesalers or manufacturers, and their record of past compliance.
The proposals were deliberately broad-brush. The current negotiations are meant to allow both sides to come to a detailed agreement on how it should all work, so long, says the EU, as it's done within the "framework" of the Protocol.
As of this week, while there has been a lot of technical work on customs, there is still no breakthrough.
The UK wants no checks on goods destined for end-consumers in Northern Ireland, the EU wants reassurance and deeper, real time access to UK databases. Even then, Brussels will not accept a blanket outsourcing to traders.
"It's about reducing the risk level to the single market, to reassure member states," says one EU diplomat.
"That involves very technical issues about access to databases, what sort of databases, what sort of access. Then there are issues about customs codes and their uses so that individual goods can be tracked."
Another big source of tension in the negotiations, both under Theresa May and Boris Johnson, was who would carry out checks and controls at Northern Irish ports.
According to the OUP commentary, the EU initially proposed they be carried out jointly, by EU and UK officials.
This was fiercely resisted by the UK as an encroachment on its sovereignty. In the event, the EU conceded that UK officials would carry out the checks, but with member state officials getting a supervisory/monitoring role.
This is covered in Article 12 of the Protocol.
The commentary says the language here was strengthened to compensate for the EU’s concession on joint controls.
Essentially, Article 12(1) means that the legal obligations that fall on member states to apply EU law also fall on the UK in implementing the Protocol at ports.
"The reason for including Article 12(1) was thus likely the wish of the Parties to prominently emphasise that the UK authorities are responsible for applying the Protocol in Northern Ireland," says Lieflӓnder, "and thereby clearly to distinguish this arrangement from the EU's initial proposal of joint controls."
The row over whether or not those EU officials doing the supervising should be allowed an office in Belfast is a reminder of how sensitive this issue was for the UK.
"Exchanges between the EU and the UK immediately following the entry into force of the Withdrawal Agreement concerning the EU’s right to be present demonstrated that the issue of the EU’s presence was sensitive, and that the Parties did not necessarily have the same interpretation of the scope of the EU’s rights and the UK’s obligations…"
Another sensitive issue is how Northern Ireland will keep up with changes in EU legislation.
As mentioned, the list of single market rules are contained in annexes attached to the Protocol - a total of 338 "acts" when the deal was concluded.
Under Article 13(3), if those rules are amended or replaced, the change in Northern Ireland is immediate.
"This process is fully automatic; it requires neither the consent of the UK nor any formal process of updating the Protocol or its annexes," says the OUP commentary.
The commentary is frank on the fact that this raises questions of democratic accountability.
"From the EU’s perspective, this set-up is crucial: the Protocol can only achieve its objectives if Northern Ireland remains aligned with relevant rules, and as those rules evolve so must Northern Ireland’s alignment therewith," writes Lieflӓnder.
"After all… Northern Ireland will be bound by developments in EU law for a potentially open-ended period without any formal 'say’ in how this law is made."
This, says Lieflӓnder, is offset by Article 15, which obliges the EU to inform the UK about all updating of single market rules and, the Article adds, "[to] ensure that all views expressed by the United Kingdom [ . . . ] are communicated to the relevant institutions, bodies, offices and agencies of the [EU] without delay".
The UK can object to a particular rule, but it has to give good grounds for doing so
It is also offset, says Lieflӓnder, at a deeper level by Article 18, which gives the Northern Ireland Assembly a consent clause, ie the vote that will take place in 2024 and periodically after that.
While the automatic update covers existing single market regulations that are simply amended, if the EU is bringing in new rules, then it must inform the UK through the Joint Committee.
Those rules are then added to the Protocol’s annexes by mutual consent (last year the Joint Committee agreed, for example, to add a new measure to reduce the impact of plastic on the environment).
The UK can object to a particular rule, but it has to give good grounds for doing so.
Rules implementing the single market are often very technical in nature and as such they are regularly updated, repealed or they expire.
A new study by research fellow Lisa Claire Whitten, of the Queen's University Belfast (QUB) Protocol unit, found that in the first year of the Protocol the number of EU "acts" listed in the annexes actually fell by 26 to 312.
Another aspect of Article 13 has been the subject of some prickly speculation by Brexiteers - the idea that the Protocol is temporary.
Under Theresa May’s deal, the "backstop" was designed to be an option of last resort, if the "alternative arrangements" or the free trade agreement did not produce the conditions that kept a hard border on the island of Ireland at bay.
Article 2 of May’s deal placed an obligation on both sides to use their "best endeavours" to supersede the Protocol through a subsequent agreement.
In Boris Johnson’s revised Protocol, there is something of a hangover of this.
Article 13 (8) says that if there is any subsequent agreement between the EU and UK it will have to indicate which parts of the Protocol it supersedes, and those parts "shall cease to apply".
"It is generally the case under public international law that two Parties to a treaty can supersede any provision of that treaty by subsequent agreement between themselves," writes Lieflӓnder.
According to Lieflӓnder, Article 13(8) symbolically signals the possibility that the Protocol can be superseded in whole or in part by a subsequent agreement, but both sides would have to be explicit as to how that agreement still meets the aims of the Protocol - avoiding a hard border on the island.
It would be "decidedly questionable", argues Lieflӓnder, whether a subsequent agreement that did not spell this out could be interpreted as "expressing the intention of the Parties to terminate, suspend, or disapply the Protocol in whole or in part."
Start pulling at threads and the thing unravels, then we're back in 2017
Thomas Lieflӓnder is a German competition lawyer who was front and centre on all Irish questions during the Protocol negotiations, and who repeatedly met Northern Irish stakeholders in Brussels and Belfast over a long period of time.
It is clear that this commentary is not an attempt to retroactively re-interpret the Protocol in light of the UK’s deepening recalcitrance on its implementation: Lieflӓnder’s contribution was concluded in December 2020, and his text - as well as those of the other contributors - was delivered to OUP at the end of January 2021.
This means that it was written before all of the controversies of last year - the Article 16-vaccines debacle, the UK unilaterally extending grace periods, the legal action that was triggered by the EU and then paused, the street disturbances, the UK Command Paper, and, of course, the turmoil of recent weeks.
It goes some way to explain why the EU is so obdurately resisting a wholesale renegotiation. Start pulling at threads and the thing unravels, then we’re back in 2017.