Induction: Complete and Incomplete Enumeration, Certainty and Probability

1. A different direction of inference

In the previous lesson, an exceptive syllogism used an established conditional and an additional premise to obtain a necessary consequence.

Induction (istiqrāʾ) works in another direction. We investigate particular cases and then draw a general claim about the class to which those cases belong.

For example, observing leaf-fall in several trees and concluding that every tree of that kind sheds its leaves is inductive. The key issue is how far the observed cases justify the general claim.

2. The source definition of induction

الاستقراء هو الحكم على كلي بتتبع أكثر الجزئيات
Induction is judging a universal by investigating most of its particular instances.

A judgment about a universal concerns a whole class, such as all animals. Investigating particulars means examining members of that class separately.

The word “most” is significant: moving from what has been examined to a claim about the entire class adds something not yet directly observed.

3. Identifying a universal and a particular

To understand induction, distinguish a universal concept from its particular instances. “Animal” may be predicated of many individual beings; one identified horse or person is a particular instance of that class.

Universal: animal.
Instances: this horse, that cow, this bird.
Observation: this particular horse has a certain feature.

Knowing one instance has a feature gives us a particular fact. Generalizing across a class from examined instances is the characteristic inductive step.

4. What investigation of cases involves

Tatabbuʿ (investigation) is not merely listing names. It means checking the relevant feature in each selected instance. If the claim concerns jaw movement during chewing, that very behavior must be examined.

Imagine a teacher testing several pens:
Pen 1 — produces ink.
Pen 2 — produces ink.
Pen 3 — produces ink.

These observations are established, but they do not yet prove that every pen of the same kind must work. The general claim requires attention to coverage and unexamined instances.

5. The textbook’s animal example

كل حيوان يحرك فكّه الأسفل عند المضغ
Every animal moves its lower jaw when chewing.

The book presents this traditional claim while illustrating incomplete induction. It mentions surveying humans, horses, cattle, donkeys, birds, and beasts.

The logical purpose is to show how a trait reported across many observed members is generalized to a whole class. The example should not be treated as a verified contemporary zoological rule.

6. Premises and conclusion in the animal example

An inductive argument has two stages: observations about individuals and a conclusion about the whole class.

Observations: human beings, horses, cattle, and other investigated animals were reported to move the lower jaw while chewing.
General conclusion: every animal moves its lower jaw in that manner.

The conclusion extends to individuals that were never examined. The central question is whether the investigated instances are sufficient to establish that extension.

7. Observation is not the universal claim

If twelve examined instances all have feature A, the statement about those twelve is supported, assuming the observations are accurate.

But “every member of this class has A” is a broader claim. It extends to members no one has examined, perhaps including future members.

Inductive reasoning must therefore assess not only whether the observations are sound but whether the scope of the generalization is justified.

8. The defining limitation of incomplete induction

In incomplete induction, the survey does not cover all the members or all the relevant cases of the class. Some remain unexamined, even when the number investigated is large.

If 100 students in a city all report studying Arabic, that does not prove that every student in the city does so. An unasked student might give a different answer.

“Incomplete” does not mean useless. It means the observations do not logically entail the universal claim with certainty.

9. The book’s conclusion: probability, not certainty

والاستقراء لا يفيد اليقين، وإنما يحصل الظن الغالب
Induction does not yield certainty; rather, it produces a prevailing probability.

The statement concerns induction that has not encompassed all particulars. When some cases remain unknown, an instance contrary to the universal claim remains possible.

Prevailing probability (ẓann ghālib) is not a baseless guess: the evidence supports the judgment, but it does not eliminate every logical possibility of an exception.

10. The logical role of the crocodile exception

إن التمساح ليس على هذه الصفة، بل يحرك فكّه الأعلى
The crocodile is not described as having that property; rather, it moves its upper jaw.

This is a traditional example presented by the book. Its function is to illustrate a possible exception: a feature found in investigated animals need not belong to every other animal.

The zoological detail need not be accepted as a verified modern fact for the logical lesson to work. A genuine counterexample would invalidate the statement that every member has the property.

11. Why one real counterexample matters

A universal affirmative claim has the form “Every A is B.” To refute it, one verified instance that is A but not B is enough.

Claim: Every student is present.
Counterexample: Ahmad is a student and is absent today.
Conclusion: “Every student is present” is false.

One confirming case does not establish a universal claim, whereas a genuine counterexample can refute it. This asymmetry is important in assessing induction.

12. Evidence can support different degrees of confidence

Not all incomplete inductions are equally persuasive. A conclusion based on two observations may be weaker than one supported by many observations across different relevant conditions.

Numbers alone, however, do not suffice. Two hundred cases drawn from one narrow subgroup may leave important cases unexamined.

The strength of the evidence may increase without the conclusion becoming a logical necessity. Full entailment calls for complete relevant coverage and true premises.

13. Complete induction

The book’s notes introduce complete induction, in which every relevant particular of the universal class is covered.

If a school has exactly four classrooms and each one is checked and found to have a fan, we may conclude that every classroom in that school has a fan.

The certainty comes from the completeness of the enumeration: there is no unexamined fifth classroom within the defined class.

14. The source on complete induction

هو إما تام إن كان حاصراً لجميع الجزئيات
Induction is complete when it encompasses all particulars.

This defines completeness by exhaustive coverage, not by the impressive size of a sample.

Examining five members can be complete when the class has only five. Examining ten thousand is still incomplete when relevant members remain outside the investigation.

15. Complete induction as a divisive syllogism

The notes also call complete induction al-qiyās al-muqassam—a syllogism by exhaustive division. The class is divided into all its relevant kinds, and the property is established for each kind.

Suppose a collection contains only categories A, B, and C, and a feature D is established for each category. Because the division is exhaustive, the whole collection has D.

Completeness of the division is part of the proof. If a relevant fourth category was missed, the conclusion no longer follows in that manner.

16. The book’s example of bodies

كل جسم إما جماد أو حيوان أو نبات
Every body is either inanimate matter, an animal, or a plant.

This is the opening premise of the traditional example of complete induction in the book’s notes. It divides the class of bodies into three stated categories.

For the logical inference, the division must cover every relevant case. This is a historical logical classification, not an independently verified modern scientific taxonomy.

17. Completing the argument about bodies

In the notes, the classes of bodies are first specified; then each class is assigned occupying space (taḥayyuz).

Premise 1: Every body belongs to one of the three classes.
Premise 2: Every member of each class occupies space.
Conclusion: Every body occupies space.

The conclusion follows within the argument only because the division is taken as exhaustive and the predicate is established for all its classes. If even one relevant class is unaccounted for, the proof is incomplete.

18. Why complete induction can be certain

The notes state: وهو يفيد اليقين
And it yields certainty.

Certainty here depends on the structure of the reasoning, not on a label: all relevant cases are covered and the property holds in every case. No member remains outside the premises to overturn the conclusion.

If the division is incomplete or a premise is false, the inference cannot be relied upon as a sound complete induction. Certainty presupposes the truth and completeness of the premises.

19. Why exhaustiveness matters

Suppose students in two classes are shown to have sat an examination. Does that prove every student in the school sat it?

Yes if these are the only two classes and the claim is established for every student in both. No if a third class exists and has not been considered.

Complete induction therefore needs two checks:
1. Does the division cover the entire class?
2. Is the property established for every member of each part?

20. Direct enumeration versus exhaustive division

Complete induction may proceed by checking each member or by dividing the whole class into exhaustive kinds.

Enumerating members: A building has exactly six known doors, all checked and found closed.
Enumerating kinds: A collection of books has exactly two sections, and every book in both has the same seal.

Both methods depend on excluding any overlooked member or relevant subdivision of the class about which the judgment is made.

21. Completeness is not a matter of large numbers

Before deciding whether an induction is complete, define the scope of the universal class and ask whether any relevant members remain unknown.

Example 1: A room contains exactly three chairs; all three are checked — complete induction.
Example 2: Three hundred chairs are checked in a market, and a claim is made about every chair in the city — incomplete induction.

The first investigation is small but exhaustive. The second is large but incomplete. “Complete” and “numerous” are not synonyms.

22. Comparing certainty and prevailing probability

Complete induction: All relevant members or kinds are covered and the property is established for each; the conclusion follows with certainty from sound premises.
Incomplete induction: Some members are known and others are not; the observations may support the conclusion, but cannot entail a universal judgment.

Verifying that all four known rooms are lit supports a certain claim about those four rooms. Seeing lights in selected homes does not justify the same level of certainty about every home in a city.

The decisive difference is coverage relative to the claim, not the raw quantity of evidence.

23. Induction compared with deduction

Deduction (qiyās): A conclusion is drawn from premises structured to entail it, often applying a universal rule to a particular case.
Induction (istiqrāʾ): A general judgment is drawn from facts established about members of a class.

Deduction:
Every human is mortal. Zayd is human. Therefore Zayd is mortal.
Induction:
All five known students in this class are present. Therefore every student in the class is present.

The inductive example is certain only if those five are in fact the entire class. An incomplete list makes the generalization incomplete.

24. Three common errors in induction

Error 1: “I have seen many cases, so every case is the same.” Many instances do not guarantee a universal conclusion.

Error 2: “My sample is large, therefore my induction is complete.” Completeness requires exhaustive coverage, not impressive numbers.

Error 3: “Incomplete induction has no intellectual value.” It may support prevailing probability; the mistake is to treat that probability as certainty.

Correct analysis defines the scope of the general claim and describes the strength of evidence accordingly.

25. Worked comparison: a class and a city

Case A: A class contains exactly eight students. Each one is checked and found present. “Every student in this class is present” is complete induction.

Case B: Eight classes in a city have full attendance. “Every class in the city has full attendance” is incomplete, because other classes were not examined.

The word “all” does not settle the issue. We must identify which entire set is under discussion and determine whether it was fully investigated.

26. Summary and the next topic

Induction draws a universal judgment from an investigation of particulars or subdivisions.

Complete induction: Exhaustive relevant coverage; true and complete premises yield certainty.
Incomplete induction: Some cases remain unknown; evidence may support a prevailing probability without necessitating the universal judgment.

The book’s animal example illustrates incomplete generalization; its traditional division of bodies shows the logical role of exhaustive classification.

The next subject is analogy (tamthīl), which extends a judgment from one particular to another through an alleged shared feature, rather than directly judging an entire class.