12 — Walā (Patronage Inheritance): The Degenerate Projection

12 — Walā (Patronage Inheritance): The Degenerate Projection

Core claim: Walā (الولاء) — the causal ʿaṣaba bond between a manumitter and a freed slave — is not a complex extension. It is a simple jiha extension ($j = 5$) combined with a pre-processing resolution function that is structurally identical to the existing priority cascade. No new axioms, no new pipeline phases, no new exceptions.

---

1. Source Text Foundations

1.1 The Three Causes of Inheritance

«أسباب الإرث ثلاثة: نكاح ونسب وولاء»

• Nikāḥ (marriage): modeled as $j = 0$ (spouse).
• Nasab (blood): modeled as $j \in {1, 2, 3, 4}$ (descendant, ascendant, sibling, uncle).
• Walā (patronage): currently unmodeled in the 5-tuple formalism.

1.2 The Walā Jiha in Classical Ordering

The source text (faraid/reference.txt) enumerates exactly four jiha for ʿaṣaba priority:

$$\text{بنوة}(1) \succ \text{أبوة}(2) \succ \text{فروع أبوة}(3{,}4) \succ \text{ولاء}(5)$$

The detailed model in faraid/asaba.md expands this to 5–7 directions depending on madhhab, but walā is universally last in all enumerations.

1.3 The Two Categories of Walā Heirs

«العصبة السببية صنفان: الأول: المعتق، سواءٌ كان ذكرًا أم أنثىٰ. الثاني: عصبة المعتق بالنفس، فلا تدخل العصبة بالغير، ولا العصبة مع الغير.» — faraid/asaba.md

Category 1: The muʿtiq (patron) themselves — male or female. Category 2: The muʿtiq’s ʿaṣaba bi-l-nafs (self-standing male agnates only). Explicitly: NOT ʿaṣaba bi-l-ghayr (female-rescue), NOT ʿaṣaba maʿa al-ghayr (sisters-with-daughters).

1.4 The Two Conditions for Walā Inheritance

«يرث المعتقُ الرقيقَ الذي أعتقه بعد موت الرقيق بشرطين: الشرط الأول: أن لا يوجد أصحاب فرض يستغرقون التركة. الشرط الثاني: أن لا توجد عصبة بالنَّسَب.» — faraid/asaba.md

Condition 1: Farḍ heirs do not exhaust the estate ($\sum \text{farḍ} < 1$). Condition 2: No nasab ʿaṣaba exists (or existing ones are impeded by a māniʿ).

1.5 Internal Ordering Within Walā

«والأقرب في جهة الولاء: المعتق ثم عصبته كترتيب عصبة النسب.» — faraid/reference.txt

Within walā, ordering is: muʿtiq first, then the muʿtiq’s male agnates using the same $(j, d, q)$ cascade — but relative to the muʿtiq, not the deceased.

Example: «أن يموت شخص عن ابن معتقه وعم معتقه. فلابن المعتق جميع المال تعصيبًا.»

(Muʿtiq’s son > muʿtiq’s uncle, because within walā the nasab ordering applies relative to the muʿtiq.)

1.6 The Non-Feminization Constraint

«المعتق وعصباته من الذكور لا يعصِّبون مَن بإزائهم من الإناث.» — faraid/asaba.md, Fāʾida

Walā ʿaṣaba do NOT create ʿaṣaba bi-l-ghayr for females. A female muʿtiq inherits directly (she herself has the walā bond), but the muʿtiq’s female relatives (daughters, sisters) cannot inherit through walā.

---

2. The Breakthrough: Walā is a Degenerate Projection

2.1 Structural Comparison with Existing Extensions

Extension	Nature	Heirs per seat	Pipeline impact
Munāsakhat (Chain)	Multi-deceased tree evaluation	Multiple (sub-estate distribution)	Multi-pass: jāmiʿa merge
Mafqūd (Missing)	Uncertainty envelope	Same heirs, two evaluations	Multi-pass: min-envelope
Dhawī al-Arḥām (Projection)	Seat → beneficiary mapping	Multiple (projection)	Multi-pass: tanzīl
Walā	Single-seat, winner-take-all	Exactly one	Pre-processing only

Walā is the simplest case of projection: one seat, one winner, no distribution.

2.2 Why Walā is Not a Full Extension

The three named extensions (Chain, Min, Project) are multi-pass structures — they require evaluating the single-pass pipeline multiple times and composing results. Walā does not:

1. It does not create a sub-problem that feeds back into the main pipeline.
2. It does not require distributing a share among multiple beneficiaries.
3. It produces exactly one heir vector that enters the standard single-pass pipeline.

Therefore, walā is a pre-processing resolution (like BFS path resolution), not a pipeline extension.

---

3. The Formal Model

3.1 Jiha Extension

Extend the jiha domain:

$$j \in {0, 1, 2, 3, 4, 5}$$

$j$	Name	Arabic
0	Spouse	زوجية
1	Descendant	بنوة
2	Ascendant	أبوة
3	Sibling	أخوة
4	Uncle	عمومة
5	Walā (Patronage)	ولاء

The ʿaṣaba priority ordering in Axiom $\alpha_2$ extends naturally:

$$\pi(\vec{h}) = (j, d, q) \quad \text{with} \quad j = 5 \text{ ranked last}$$

3.2 The Walā Resolution Function

Definition. Given a muʿtiq $m$ (patron) and a family state, the walā resolution function $\mathcal{R}_W$ selects the actual walā heir:

$$\mathcal{R}W(m) = \begin{cases} m & \text{if } m \text{ is alive and eligible (no impediments)} \[4pt] \displaystyle\operatorname{argmin}{\pi}!\left(A_m^{\text{nafs}}\right) & \text{if } m \text{ is dead or impeded} \end{cases}$$

where:

• $A_m^{\text{nafs}} = {h : h \text{ is a male agnate (ʿaṣaba bi-l-nafs) of } m, \text{ alive and eligible}}$
• $\pi(h) = (j_h^{(m)},, d_h^{(m)},, q_h^{(m)})$ is the standard priority ordering evaluated relative to $m$ (not the original deceased)
• $\operatorname{argmin}_{\pi}$ selects the highest-priority element (lowest lexicographic value)

Key property: This resolution uses the same $\alpha_2$ cascade that governs nasab ʿaṣaba — a recursive self-application of the formalism.

3.3 The Walā Heir Vector

The resolved walā heir $w = \mathcal{R}_W(m)$ enters the freed person’s pipeline as:

$$\vec{w} = (g_w, ; 5, ; 1, ; 9, ; 1)$$

Component	Value	Rationale
$g_w$	Gender of $w$	Male (if from $A_m^{\text{nafs}}$) or M/F (if $w = m$, since muʿtiq can be female)
$j = 5$	Walā jiha	Lowest ʿaṣaba priority
$d = 1$	Degree 1	Direct walā bond (no generational chain in the patronage relationship)
$q = 9$	N/A	Quwwa is structurally inapplicable to walā (no full/paternal/maternal distinction)
$c = 1$	Valid	The resolution function guarantees eligibility

3.4 Why $d = 1$ and $q = 9$ Are Correct

• Degree: The walā bond is a direct contractual relationship — there is no “generational distance” between patron and freed person. Unlike $j = 0$ (spouse, $d = 0$), we use $d = 1$ because the walā heir is a single person who takes the residual (not a special farḍ class). The exact value is irrelevant since there is at most one walā heir — no $d$-based tiebreaking ever occurs within $j = 5$.

• Quwwa: The source text restricts quwwa to collateral heirs: «ولا يتصور التقديم بالقوة إلا في جهة فروع الأبوة». Walā has no full/paternal/maternal subdivision, so $q = 9$ (N/A) is the honest encoding.

3.5 Uniqueness of the Walā Heir

Theorem. In any single inheritance problem, at most one walā heir enters the pipeline.

Proof: The classical rule «إنما الولاء لمن أعتق» assigns walā to exactly one patron per freed person. The resolution function $\mathcal{R}_W$ returns exactly one person (or $\emptyset$ if no eligible heir exists in the muʿtiq’s ʿaṣaba hierarchy). Therefore $|W| \le 1$.

Corollary: No intra-walā competition (by $d$ or $q$) ever occurs within the freed person’s pipeline. All competition happens inside the resolution function, which operates in the muʿtiq’s reference frame.

---

4. Integration with the Axiom System

4.1 No New Axioms Required

Axiom	Handles walā?	How?
$\alpha_1$ (Intermediary Exclusion)	✓	The walā heir is not anyone’s intermediary in the freed person’s kinship DAG.
$\alpha_2$ (ʿAṣaba Ranking)	✓	$j = 5$ loses to $j \in {1, 2, 3, 4}$ — Condition 2 is automatically satisfied.
$\beta$ (Entitlement)	✓	Walā heir inherits as ʿaṣaba (no farḍ share); standard residual absorption applies.
$\gamma_1$ (Residual Absorption)	✓	If $\sum \text{farḍ} < 1$ and no nasab ʿaṣaba exists, walā heir absorbs remainder — Condition 1 satisfied.
$\gamma_2$ (ʿAwl)	✓	If $\sum \text{farḍ} \ge 1$, walā heir gets 0 (same as any ʿaṣaba).
$\gamma_3$ (Radd)	✓	Radd occurs when no ʿaṣaba exists. If walā heir exists → they are ʿaṣaba → radd doesn’t trigger.
$\delta$ (Kinship Monotonicity)	N/A	$\delta$ addresses blood-path supersets. Walā is not a blood relationship; $\delta$ is inapplicable.

Both conditions from §1.4 emerge as natural consequences of $\alpha_2$ and $\gamma$:

• Condition 1 ($\sum \text{farḍ} < 1$) = standard ʿaṣaba precondition under $\gamma_1$.
• Condition 2 (no nasab ʿaṣaba) = $\alpha_2$ priority: $j \in {1..4}$ always outranks $j = 5$.

4.2 The Non-Feminization Constraint is Vacuously Satisfied

The pipeline’s ʿaṣaba bi-l-ghayr mechanism triggers when: a male ʿaṣaba at $(g=1, j, d, q)$ coexists with a female at $(g=0, j, d, q)$ (same class). For walā:

• At most one walā heir enters the pipeline (Theorem, §3.5).
• There is no female walā heir at the “same class” (the female muʿtiq is either the resolved heir herself, or absent).
• Therefore, the 2:1 rescue mechanism cannot fire for $j = 5$.

The source text’s explicit constraint (“المعتق وعصباته لا يعصِّبون مَن بإزائهم من الإناث”) is automatically satisfied by the resolution function’s restriction to ʿaṣaba bi-l-nafs. The pipeline needs no special case.

---

5. The Resolution Function in Detail

5.1 The Recursive Structure

$$\mathcal{R}W(m) = \begin{cases} m & \text{alive}(m) \wedge \text{eligible}(m) \[3pt] \text{Top}\left(\text{BFS}(m) \big|{\text{nafs}}\right) & \text{otherwise} \end{cases}$$

Where $\text{BFS}(m)\big|_{\text{nafs}}$ means: run the standard BFS resolver with $m$ as the “deceased,” then filter to only ʿaṣaba bi-l-nafs candidates:

$$\text{Filter} = {h : c(h, m) = 1 ;\wedge; h \text{ is ʿaṣaba bi-l-nafs of } m}$$

The ʿaṣaba bi-l-nafs constraint means:

• $j_h^{(m)} \in {1, 2, 3, 4}$ (not spouse)
• The heir must be male ($g_h = 1$), except if the heir is the son (who is always ʿaṣaba bi-l-nafs), or specifically a male in a category that IS ʿaṣaba bi-l-nafs.

More precisely, ʿaṣaba bi-l-nafs of $m$ are exactly:

1. Sons ($j=1, d=1, g=1$)
2. Sons’ sons and below ($j=1, d \ge 2, g=1, c=1$)
3. Father ($j=2, d=1, g=1$)
4. Grandfather and above ($j=2, d \ge 2, g=1, c=1$)
5. Full/paternal brothers ($j=3, d=1, q \in {1,2}, g=1$)
6. Brothers’ sons and below ($j=3, d \ge 2, q \in {1,2}, g=1, c=1$)
7. Full/paternal uncles ($j=4, d \ge 1, q \in {1,2}, g=1$)
8. Uncles’ sons and below ($j=4, d \ge 2, q \in {1,2}, g=1, c=1$)

This is the standard male agnate hierarchy — the same set that inherits in any standard problem under the ʿaṣaba bi-l-nafs category.

5.2 Ordering Within the Resolution

The ordering of the muʿtiq’s ʿaṣaba follows the source text:

«أن يموت شخص عن ابن معتقه وعم معتقه. فلابن المعتق جميع المال تعصيبًا.» (Muʿtiq’s son beats muʿtiq’s uncle.)

«أن يموت شخص عن ابن ابن ابن أخي معتقه وعم معتقه. فلابن ابن ابن أخي المعتق جميع المال تعصيبًا.» (Muʿtiq’s brother’s great-grandson beats muʿtiq’s uncle — because brother’s line precedes uncle’s line.)

This is exactly the standard $\alpha_2$ cascade: jiha → daraja → quwwa, applied in the muʿtiq’s reference frame.

5.3 Edge Cases

Case A: Female muʿtiq alive. $\mathcal{R}_W(m) = m$ (the female patron). Vector: $(0, 5, 1, 9, 1)$. She takes the residual.

Case B: Female muʿtiq dead, her son alive. $\mathcal{R}_W(m) = \text{son}(m)$. The son is ʿaṣaba bi-l-nafs of $m$. Vector: $(1, 5, 1, 9, 1)$.

Case C: Muʿtiq dead, no eligible ʿaṣaba bi-l-nafs. $\mathcal{R}_W(m) = \emptyset$. No walā heir enters the pipeline. Under Radd configuration, remainder goes to blood farḍ heirs.

Case D: Muʿtiq alive but impeded (e.g., different religion, killer). $\mathcal{R}_W(m) \ne m$. Fall through to muʿtiq’s ʿaṣaba. The muʿtiq is treated as dead for walā purposes.

---

6. Worked Examples (from Source Text)

Example 1: Muʿtiq takes residual

Share	Heir	Notes
$\frac{1}{8}$	Wife	Farḍ (pressed by descendant)
$\frac{1}{2}$	Daughter	Farḍ (solo)
$\frac{3}{8}$	Muʿtiq	ʿAṣaba — walā ($j = 5$) takes remainder

Conditions: $\sum \text{farḍ} = \frac{5}{8} < 1$ ✓, no nasab ʿaṣaba ✓.

Example 2: Walā excluded by nasab ʿaṣaba

Share	Heir	Notes
$\frac{1}{6}$	Mother	Farḍ (pressed by descendant)
Remainder	Son	ʿAṣaba — nasab ($j = 1$)
0	Muʿtiq	Blocked: nasab ʿaṣaba exists ($j = 1 \prec j = 5$)

The $\alpha_2$ ranking handles this automatically.

Example 3: Walā excluded by farḍ exhaustion

Share	Heir	Notes
$\frac{1}{2}$	Husband	Farḍ
$\frac{1}{2}$	Full Sister	Farḍ (solo)
0	Muʿtiq	Blocked: $\sum \text{farḍ} = 1$ (no remainder for ʿaṣaba)

Under $\gamma_1$: ʿaṣaba gets $1 - \sum \text{farḍ} = 0$.

Example 4: Nasab ʿaṣaba impeded, walā inherits

Share	Heir	Notes
$\frac{1}{2}$	Husband	Farḍ
$\frac{1}{6}$	Mother	Farḍ (pressed by siblings ≥ 2 — ghost pressure from blocked brother)
0	Full Brother (killer)	Impeded by māniʿ → excluded in Phase 0
$\frac{1}{3}$	Muʿtiq	ʿAṣaba — walā takes remainder (nasab ʿaṣaba exists but is impeded)

Example 5: Muʿtiq dead, their son inherits

Deceased (freed person) leaves: mother + muʿtiq’s son + muʿtiq’s uncle.

Resolution: $\mathcal{R}_W(m) = \text{son}(m)$ (son’s $\pi = (1, 1, 9)$ beats uncle’s $\pi = (4, 1, 1)$).

Share	Heir	Notes
$\frac{1}{3}$	Mother	Farḍ (no pressors)
$\frac{2}{3}$	Muʿtiq’s son	ʿAṣaba — resolved walā heir
—	Muʿtiq’s uncle	Not selected (lost in walā resolution)

---

7. The Bayt al-Māl Extension ($j = 6$)

For completeness, the Mālikī/Shāfiʿī position adds a 7th jiha — the public treasury (بيت المال) as a final residual absorber:

$$j = 6: \text{Bayt al-Māl (Treasury)}$$

Under this configuration, if no walā heir exists and $\sum \text{farḍ} < 1$:

• The treasury absorbs the remainder (instead of Radd expanding to blood farḍ heirs).
• Mathematically: a phantom heir vector $(1, 6, 1, 9, 1)$ that always exists.

This is controlled by the dispute toggle D3 in findings/10-dispute-matrix.md:

• Radd configuration (Ḥanbalī, Path B): $j = 6$ absent, $\gamma_3$ handles excess.
• Treasury configuration (Mālikī/Shāfiʿī): $j = 6$ present, pre-empts $\gamma_3$.

---

8. Summary: What Walā Adds to the Formalism

Layer	Change	Magnitude
5-Tuple	$j$ domain extends from ${0..4}$ to ${0..5}$ (or ${0..6}$ with treasury)	1 enum value
Axioms	None. $\alpha_2$ and $\gamma_1$ already handle walā’s behavior.	Zero
Exceptions	None. The non-feminization constraint is vacuously satisfied.	Zero
Pipeline	No new phases. Walā resolution is a pre-processing step (like BFS).	Zero
Pre-processing	One new function: $\mathcal{R}_W(m)$ — resolve the walā heir.	1 function
DAG	One new edge type: walā bond (patron → freed person).	1 edge type

8.1 The Architectural Elegance

The walā mechanism reuses the same mathematical structure at two levels:

1. Outer level (freed person’s pipeline): The resolved walā heir competes by $\alpha_2$ ranking at $j = 5$.
2. Inner level (walā resolution): The muʿtiq’s agnates compete by $\alpha_2$ ranking at $j \in {1..4}$.

This is a composition of the priority cascade with itself — formally:

$$\text{walā_heir}(\text{freed_person}) = \alpha_2\text{-best}\left(\text{ʿaṣaba-bi-l-nafs}(\text{muʿtiq})\right) \quad \text{at} \quad j = 5$$

The formalism is self-similar across reference frames.

---

9. Implementation Notes for the Engineer

1. DAG edge type: Add a wala edge from muʿtiq → freed person (in addition to parent and spouse edges). This is a non-blood, non-marital directed edge.

2. Resolution step: Before computing the pipeline for a freed person, check if a wala edge exists. If so, resolve $\mathcal{R}_W(m)$:

   • If $m$ alive and eligible → walā heir is $m$.
   • Else → run BFS from $m$‘s perspective, filter to ʿaṣaba bi-l-nafs, select top by $\pi$.

3. Vector injection: The resolved heir enters the pipeline as $(g_w, 5, 1, 9, 1)$.

4. No special cases in Phase 2/3/4: The existing $\alpha_2$ ranking and $\gamma_1$ absorption handle everything. No new conditionals needed inside the pipeline.

5. Testing strategy:

   • Case: farḍ heirs + walā heir → walā gets remainder.
   • Case: nasab ʿaṣaba exists → walā blocked.
   • Case: farḍ exhausts estate → walā gets 0.
   • Case: muʿtiq dead → resolution selects their top male agnate.
   • Case: female muʿtiq alive → she inherits directly.
   • Case: no muʿtiq ʿaṣaba → walā seat empty → radd applies.

---

10. Open Questions

Q8: Does Walā Chain?

If the freed person (A) themselves freed another slave (B), does A’s patron (M) inherit from B through A? The classical rule is «الولاء لا يورث وإنما يُورث به» — walā is not inheritable property, but one inherits BY it. This suggests walā does NOT chain transitively.

Tentative answer: No chaining. Each freed person has exactly one walā edge to their direct patron. Multi-level patronage (patron’s patron) is not a valid inheritance path.

Q9: Joint Muʿtiq

If two people jointly freed a slave, how is the walā distributed? Classical texts suggest each patron gets a proportional share of the walā. This would mean multiple walā heirs in a single problem, breaking the uniqueness theorem (§3.5).

This requires further source text research. If confirmed, the model extends to: multiple walā heirs at $j = 5$, each with a fractional “walā weight,” sharing the residual proportionally. This would still not require new axioms — just weighted ʿaṣaba absorption.

---

References

• Walā conditions and categories: faraid/asaba.md §القسم الثاني: العصبة السببية
• Jiha ordering with walā: faraid/reference.txt §ترتيب العصبة
• Internal walā ordering examples: faraid/reference.txt (ابن معتقه vs. عم معتقه)
• Non-feminization: faraid/asaba.md فائدة (line 281)
• Existing 5-tuple formalism: findings/01-5tuple-and-graph.md
• Axiom system: findings/02-axioms.md
• Dispute matrix (D14, jiha count): findings/10-dispute-matrix.md