Building an agreement to survive its own reversal

The situation

'If we try living together again and it doesn't work, does everything we're about to sign just disappear?' That was the question Cristina asked before she and Shira had even finished negotiating their separation agreement, and it turned out to be the most important question in the file, one that shaped nearly every clause that followed it.

Cristina and Shira had been married for close to two decades. Cristina worked as a court clerk, a steady job with a modest but reliable income, and Shira worked as a surveyor, often travelling to sites around the region for days at a stretch. Their son Yael, now an adult, had a disability that meant he would likely need some level of support and supervision for the rest of his life. Cristina had built her schedule around his needs for years, arranging her shifts to be home when he needed her and coordinating the various supports, from appointments to routine daily care, that kept his life stable.

The separation itself was not dramatic. Cristina and Shira had simply grown apart, and after months of living increasingly separate lives under one roof, they agreed to formalize the split. Negotiations over the following weeks worked through the usual questions of property, the family home, and each spouse's pension, along with a more complicated set of terms addressing how they would continue sharing responsibility for Yael's care and the costs that came with it, since he was not going to become financially independent the way most adult children eventually do, and both parents wanted that arrangement to be dependable regardless of what happened between the two of them personally.

Partway through drafting, Cristina raised the possibility that she and Shira might try to reconcile before finalizing anything. They had talked, tentatively, about whether ending the marriage was really the right call, and neither of them wanted to rule out trying again. But Cristina had heard, vaguely and secondhand, that getting back together after signing a separation agreement could undo it entirely, and she did not want to sign something that would leave her and Yael unprotected if a reconciliation attempt failed. She wanted an agreement that worked whether they got back together or not, and she wanted an honest answer before she signed anything, not after.

The problem

Cristina's instinct was not wrong. A separation agreement is available to spouses who are living separate and apart — the idea that there must be no reasonable prospect of resuming the relationship belongs to fixing the valuation date for property and to the ground for divorce, not to signing the agreement itself. But if spouses genuinely reconcile and resume living together as a couple for a meaningful period, that separation is undone, and provisions in the agreement, particularly ones tied to the fact of separation, can become vulnerable to challenge if the relationship later ends again. An agreement silent on this risk leaves the question to be argued out after the fact, at exactly the moment when the parties are least able to agree calmly on anything, having just been through a second breakup on top of the first.

The stakes were higher than they would be for most couples, because of Yael. The agreement's terms addressing his ongoing support were not a standard child support calculation with an obvious end date attached to an age or a graduation. They anticipated years, potentially decades, of shared responsibility for an adult who was not going to age out of needing help in the way most children eventually do. If a short reconciliation attempt threw the whole agreement into question, Cristina risked having to renegotiate those terms from scratch at a later and possibly more difficult moment, with no guarantee Shira would agree to the same framework a second time around, particularly if the second separation was more acrimonious than the first.

There was also a practical timing problem. Cristina and Shira wanted to try reconciling relatively soon, potentially within weeks of signing, which meant the agreement had to anticipate a specific, near-term scenario rather than a vague possibility years down the road. Most separation agreements do not address resumption of cohabitation in any detail, because most couples who sign one are not actively planning to test the relationship again shortly afterward. This file needed a solution built for that exact situation, not a generic boilerplate clause borrowed from a template that assumed reconciliation was purely hypothetical and unlikely to actually happen.

The core problem, then, was drafting an agreement flexible enough to survive a genuine, good-faith attempt at reconciliation without collapsing, while still being firm enough to function as a real, binding separation agreement if the attempt succeeded and the parties wanted to formally set it aside. Those two goals pull against each other, and getting the balance wrong in either direction would have left Cristina exposed, either to a reconciliation attempt she felt she could not actually try, or to an agreement with no real force once she tried it.

What we did

1. Built a resumption-of-cohabitation clause into the agreement's core structure, rather than treating it as an afterthought or a single throwaway line, specifying that the parties could live together again for a defined trial period without automatically voiding the agreement. Without that clause, any cohabitation at all could later have been used to argue the agreement's separation premise no longer held, forcing the question into court at the worst possible moment; building it in up front gave Cristina and Shira room to genuinely try reconciling without legal risk hanging over the attempt.
2. Set a clear time threshold in the clause, stating that cohabitation beyond a defined number of months would be treated as a genuine resumption of the relationship requiring the parties to either formally set the agreement aside or renegotiate it, which meant a short attempt that failed would leave the original agreement fully intact and enforceable without either party needing to prove anything about their intentions.
3. Separated the property and support terms from the reconciliation clause, so a short-term reunion would not automatically reopen questions about the house, the pensions, or Yael's support arrangements. Tying every clause to the same reconciliation trigger would have meant one shared threshold governing issues that had nothing to do with each other, risking the whole negotiated package unraveling over a single ambiguous fact; keeping them structurally separate protected months of careful negotiation from being undone by a change nobody actually wanted to have to revisit from the beginning.
4. Drafted Yael's support terms to stand independently of the couple's relationship status, framing them around his ongoing needs rather than around the fact of the parties' separation, so that even if a court were later asked to look at whether reconciliation affected the agreement, the provisions built around his care were less exposed than terms tied more directly to the separation itself.
5. Documented the parties' shared intent in the agreement's recitals, recording plainly that both Cristina and Shira understood the agreement to remain the operative arrangement unless a longer reconciliation triggered the threshold. This mattered because recitals capturing a couple's stated understanding at the time of signing carry real weight if anyone later tries to argue the agreement meant something different, and it removed ambiguity about what either of them believed they were agreeing to, rather than leaving that understanding to memory.
6. Reviewed the clause specifically against Yael's situation, checking line by line that nothing in the resumption-of-cohabitation language could be read as making his support terms conditional on the couple's relationship status. A general clause drafted for a typical couple could easily have created an unintended gap for a dependent adult with ongoing needs, so this review confirmed his care arrangements would keep running on their own terms even if the reconciliation attempt dragged on longer than expected or ended more messily than either parent hoped.
7. Advised Cristina to keep records during any reconciliation attempt, suggesting she note the date they resumed living together and keep basic evidence of how the household actually functioned, from shared expenses to who was sleeping where. Memory alone tends to blur once a second separation gets emotional, and clear contemporaneous notes meant that if the timeline ever became relevant to whether the clause's threshold had been crossed, Cristina would have something more reliable than recollection to point to.
8. Encouraged both parties to sign promptly once the terms were settled, rather than letting the agreement sit unsigned while the reconciliation conversation continued in the background. An agreement left open for further discussion tends to get reopened, term by term, as second thoughts creep in, and the longer it sat unsigned the more likely the carefully balanced resumption-of-cohabitation clause was to get diluted; signing while both parties still agreed locked the protection in before anyone's thinking could drift.

The outcome

Cristina and Shira did try reconciling, moving back in together within about six weeks of signing the agreement. It did not work. Within four months, well inside the trial period the agreement specified, they concluded the relationship was not salvageable and separated again, this time for good. Because the resumption-of-cohabitation clause had anticipated exactly this scenario, the original agreement simply resumed operation without any need to renegotiate a single term, and without either of them needing to argue about what living together again had or had not meant.

The early tactical decision that ended up mattering most was Shira's, made without any apparent thought to its later effect. During the negotiation stage, before the reconciliation attempt was even on the table, Shira's side had pushed to finalize and sign the agreement quickly rather than delay for further back-and-forth over minor wording, largely to avoid additional legal fees. That urgency meant the resumption-of-cohabitation clause was locked in as drafted, with its specific time threshold, well before either party knew they would actually test it. Had the signing dragged on and that clause been reopened or diluted during a later, calmer conversation, the couple might easily have ended up with vaguer language that would not have held up as cleanly once the reconciliation failed, leaving Cristina to argue the point from a much weaker position four months later.

The result for Cristina was as clean as this kind of case gets. Nothing had to be renegotiated, no term was challenged, and the support arrangements protecting Yael's care carried on exactly as written, without interruption or dispute. The house and pension terms took effect on schedule as though the reconciliation attempt had never happened, because on paper, it functionally had not. The agreement did precisely what it was built to do: it let two people take a real, good-faith run at repairing their marriage without gambling their son's stability on the outcome.

What you can learn from this