Design rationale
The FAQ answers the objections people raise first. This page goes a layer deeper — the reasoning behind the draft's most structurally significant, and most likely to be challenged, choices. Each entry names the tension being resolved, the choice made, and what was deliberately given up to make it. None of these are presented as beyond question.
Why three caucuses, not two
The third caucus gives minority communities a distinct place in the Assembly alongside the Israeli and Palestinian caucuses, with equal authorized seats for all three. The aim is to ensure that communities outside the two larger national groupings have representation and a role in legislative coalitions. Caucuses organize candidates, representatives, and seats; they do not divide voters into separate electorates. Every eligible district voter receives all three caucus ballots.
Under Article 4 and Article 20, ordinary legislation needs at least one-third of the authorized membership of each caucus and a majority of the full authorized Assembly membership on the same vote. This requires support across all three caucuses, but it does not give a majority of each caucus a veto: a law can pass despite opposition from most representatives in one caucus if the required affirmative votes are secured. The threshold cannot guarantee that every community considers the outcome in its interests. Equal rights and the prohibition of domination remain separate constitutional safeguards. Parties and policy coalitions may cross caucus lines, and members of a caucus are not required to agree politically.
Why the justice system is split into three coordinate institutions
Article 10 splits what most legal systems treat as one thing — "the courts" — into three separate, non-subordinate institutions: the Judiciary (which judges), the Investigative Magistracy (which investigates), and the Court Enforcement Corps (which enforces rulings). The reasoning is about capture, not efficiency. A justice system with a single point of control is a single point of failure: if one office both investigates and judges, or if the body that enforces rulings answers to the same government it's supposed to hold accountable, the whole system can be pressured or corrupted by controlling that one office. Splitting these functions into genuinely independent, coordinate bodies — none of them subordinate to the others — means bad-faith interference in one doesn't compromise the whole system. It costs some efficiency. We think that's worth it for the accountability it buys, especially in a state built by two communities who each have real historical reasons not to fully trust the other's control over state power.
Why there's a middle tier between "constitution" and "ordinary law"
Related: the five proposed protected organic laws and their controlling status in Article 20.
Protected organic laws allow important implementation details to evolve while requiring more agreement than ordinary legislation. Under Article 20, enactment, amendment, or repeal requires a majority of the authorized membership of each Assembly caucus. Vacancies, absences, abstentions, and disqualifications do not lower that threshold. A law has this status only where the Constitution expressly designates it as protected.
These laws remain subordinate to the Constitution. They may implement its guarantees but cannot diminish them, silently amend constitutional text, or evade a higher voting requirement. They are subject to Constitutional Court review when challenged, rather than mandatory advance judicial approval. The protection comes from the higher legislative threshold, constitutional limits, and review of challenges. This leaves room to revise implementing rules through experience without making every detail constitutional text. The five supporting organic-law proposals are not enacted merely because the Constitution designates their functions.
Why enduring guarantees protect principles while allowing institutional change
Article 12 protects enduring principles such as equal dignity, democratic civilian government, equal belonging, non-domination, and judicial independence. It also protects the essential content of constitutional rights. The aim is to prevent constitutional change from being used to remove those protections while preserving the ability to correct wording and improve institutions.
The current draft does not establish an endlessly expanding list of rules that can be added to but never changed. Particular institutional forms may be replaced when their successors provide substantially equivalent or stronger protection in practical operation. Assembly size, district numbers, institutional names, voting methods, and administrative features are not immutable unless changing them would substantially impair a protected principle. Statutory rights and administrative practices do not automatically become constitutional or immutable. Constitutional changes must follow Article 12's applicable procedures and preserve the enduring guarantees; changing the method cannot serve as a way to remove the protection.
Why the transition protects rights while allowing institutional change
Related: delegated transition planning and the unresolved founding Ratification Protocol.
A staged transition allows institutions and public confidence to develop over time. Transition Section 3 protects rights, equal security, and protection against domination throughout that process. Those protections may not regress, but the draft does not freeze every temporary institution, implementation method, or statutory expansion.
Implementation methods and temporary institutions may change when their replacements provide substantially equivalent or stronger protection in practical operation and are not used to delay transition in bad faith. The test is what people remain protected by in practice, not whether an institutional arrangement has stayed identical. Statutory or administrative expansion does not automatically become constitutional or immutable. This allows institutions to adapt while preventing transition from becoming a reason to withdraw rights or expose a community to domination. Section 4 attributes bad-faith conduct to responsible actors and institutions, prohibits collective punishment, and keeps corrective action and judicial remedies within their constitutional limits.
Why criminal proceedings are inquisitorial, not adversarial
Article 10, Section 5 makes the Investigative Magistracy actively responsible for seeking the truth of a matter, rather than having two opposing sides present competing cases for a judge or jury to referee. This is a deliberate break from the adversarial model most common-law systems use. The reasoning goes back to the document's foundational principle that the justice system exists to serve truth and justice above all other considerations — and an adversarial system, by design, doesn't actually optimize for truth. It optimizes for which side argues better, which side can afford stronger representation, and which side benefits from a favorable procedural ruling. Those pressures don't disappear just because both sides mean well; they're built into the structure. An inquisitorial model, with a magistrate whose job is specifically to investigate rather than advocate, aims the system's incentives at the actual question — what happened — instead of at who wins. The tradeoff is real: this asks a lot of the integrity and competence of the Investigative Magistracy, which is exactly why that institution is built as independent, coordinate with the judiciary rather than subordinate to it, and insulated from executive pressure (Article 10, Section 3). The safeguard for the model's biggest risk is built into the model itself.
Why elections use Ranked Pairs, not party lists or first-past-the-post
Related: the proposed Electoral Law develops the algorithm below the constitutional level. The current voting rules are in Article 13.
Under Article 13, Sections 2–4, each geographic district elects one representative in each of three caucus contests. Every eligible voter receives all three ballots and may rank candidates separately in the Israeli, Palestinian, and Minority contests. A voter can leave any contest blank without affecting the others. No voter is registered to a caucus, and participation is not restricted to people identifying with the community represented by that seat. The separate eligibility and confidential community-support requirements apply to candidates.
Each contest uses Ranked Pairs. Rankings allow voters to express preferences beyond a single first choice, and pairwise comparisons inform the result. The proposed algorithm orders pairwise results by strength and locks them unless doing so would create a cycle. The design keeps the choice among individual candidates in the hands of district voters, while allowing preferences across candidates to count. The tradeoff is a more complex tally than a simple plurality count. Detailed tallying, unranked-candidate treatment, and tie handling belong to the protected Organic Electoral Law; remaining technical choices cannot restrict equal access to all three contests or turn optional rankings into compulsory participation.
Why the Transition Oversight Council needs two different vote thresholds
The Council that monitors transition and certifies benchmarks has to be powerful enough to actually hold predecessor-state actors accountable, but not so powerful that it becomes an unaccountable fourth branch of government during the most fragile period of the whole project. Transition Section 2 splits the difference by splitting the vote. Fifteen voting members, five each selected through the Israeli, Palestinian, and minority-community processes (¶10), with routine administrative action requiring only a majority overall and one-third from each component — eight overall, two from each five-member component (¶12). But the decisions that actually matter — stage certification, an adverse bad-faith finding, a corrective requirement, certification of transition completion, or changing an objective benchmark — require two-thirds overall and a majority from each component: ten overall, three from each (¶13). No consequential decision may be treated as routine administration. The effect: no single community's delegation, including the largest, can push through a finding against another community by itself, but a real accountability finding still needs broad cross-community agreement to succeed. The tradeoff is real — bad-faith conduct that only draws opposition from within one community's own delegation could escape a formal finding. The Council is also deliberately built to end itself: completion follows objective conditions in Section 8 and dissolves the Council automatically (¶16) — it isn't designed to entrench itself.
Why war authorization expires and has to be actively renewed, not just declared
A state built by communities with real, recent reasons not to trust each other's control over military force can't hand war powers to whoever holds the Prime Minister's office and call it settled. Article 11, Section 7 draws a hard line between immediate defense and everything else. Repelling an actual or imminent attack is permitted immediately under civilian command, with notice to the Speaker and caucus leadership within 24 hours and public disclosure of the action's existence, legal basis, and scope within 72 hours (¶21) — genuinely operational details can stay protected, but the fact and legal basis of the action cannot. Anything beyond that — continued, offensive, or non-immediate action — needs Assembly authorization within seven days, by a majority of each caucus, capped at sixty days, renewable only by the same cross-caucus vote, with public reporting at least every fourteen days (¶22–23). Secret war authorization is expressly prohibited. This makes sustained military action deliberately hard to wage without broad cross-caucus buy-in, which is the point: it's meant to prevent one caucus's leadership from dragging the state into open-ended conflict the other caucuses never approved. The cost is real operational friction exactly when speed often matters most — the immediate-defense carve-out is the release valve for that, not a loophole around the rest.
Why sacred-site decisions need both a Council supermajority and the affected community's own approval
The three-caucus problem has a parallel version for faith communities: a body making binding decisions about shared holy sites can't just be a majority-vote council, or the largest faiths could outvote the community whose site is actually at stake. Article 19, Section 2 answers with a dual test. Ordinary Council action already requires both two-thirds of the full Council and support from a majority of represented faith-community delegations (¶9) — not just one or the other. For anything more consequential — a permanent rule materially changing worship, stewardship, or access at a shared site — it additionally requires majority approval from each specifically affected delegation (¶10), so the community whose site is actually changing gets a real say, not a vote diluted among every faith on the Council. A narrow emergency-safety exception exists for genuine, time-boxed risk to life or the site itself, but it doesn't waive the underlying vote and stays subject to immediate judicial review (¶11). The tradeoff: even urgently-needed changes to a shared site can move slowly when several communities are affected. That's deliberate — this is exactly the category of decision the draft doesn't think a fast majority should be able to force through.