Suppose a measurement g(B) includes contributions from every subset A of B. To recover the exact contribution attached only to B, subtract smaller contributions with carefully chosen signs. On a three-element subset lattice the signs alternate by the number of removed elements. On a general partial order those coefficients need not be just plus or minus one; they are determined by the interval structure.
Predict
Predict the coefficient for the endpoints of the chain 0< 1< 2. The recurrence gives μ(0, 0)=1 and μ(0, 1)=−1, so μ(0, 2)=0. There is no direct endpoint correction after the intermediate cancellation. This contrasts with a Boolean interval of rank two, whose two distinct middle elements produce coefficient plus one. The number and arrangement of intermediate objects matter.
Experiment
Compute two small intervals.
Interval
Diagonal coefficients
First-step coefficients
Endpoint coefficient
Three-element chain
1
−1
0
Two-element subset lattice
1
−1 on two branches
1
Choose arbitrary contributions on the four subset-lattice elements, form all cumulative sums, and recover the original top contribution. This numerical check makes the cancellation visible before the general proof.
Understand
For a finite poset P define μ(x, x)=1 and μ(x, y)=−Σ over x≤z< y of μ(x, z) when x< y; use zero for incomparable pairs. Order the elements by a linear extension, meaning an ordering consistent with the partial order. The zeta matrix Z has entry one when x≤y and zero otherwise; it is upper triangular with diagonal one. The recursive μ matrix M satisfies MZ=I. Since finite square matrices with a one-sided inverse are invertible, also ZM=I. This supplies both interval cancellation identities and avoids silently assuming that the two possible sums are the same definition.
Apply
If g(y)=Σ over x≤y of f(x), then Σ over x≤y of μ(x, y)g(x)=Σ over z≤y of f(z)Σ over z≤x≤y of μ(x, y). The inner sum is zero unless z=y, when it is one, proving inversion. For subsets, the interval from A to B is a Boolean lattice on B minus A. The formula μ(A, B)=(−1)^r satisfies the recurrence because the total interval sum is (1−1)^r. Thus ordinary inclusion–exclusion is Möbius inversion on subsets. For divisibility, the interval from a to b is the divisor lattice of b/a, recovering the arithmetic Möbius function.
For a practical inversion test, use the chain with cumulative values g(0)=2, g(1)=7, g(2)=11. The recovered contributions are 2, 5, 4. At the last position, the endpoint coefficient zero means f(2)=g(2)−g(1), with no extra g(0) term. Comparing this with the Boolean-square subtraction prevents importing the subset sign pattern into an unrelated order.
Continue
Use this method when data are aggregated over a partial order and the desired values are exact rather than cumulative. Write the order relation explicitly: inclusion, divisibility and refinement are different posets with different coefficients. The proof here assumes a finite poset; locally finite infinite posets can be handled interval by interval, but arbitrary infinite summations require additional care. A linear extension is bookkeeping for the matrix proof, not an assertion that incomparable elements become comparable in the original order.
Drafts save locally as you type. Reveal worked solutions when useful; the site remembers that exposure. Automatic matches check the requested answer, while written arguments require review against the rubric.
Find μ(empty set,{1, 2, 3}) in the subset poset.
Integers, fractions and finite decimals are accepted. Expressions such as sqrt(2) are not parsed by this checker.
Find μ(0, 2) in the three-element chain 0<1<2.
Integers, fractions and finite decimals are accepted. Expressions such as sqrt(2) are not parsed by this checker.
Prove Σ over x≤z≤y of μ(x, z) is 0 for x<y.
Structure your argument and preserve revisions
These optional planning fields save locally. Recording your response saves an immutable copy of the argument and these fields. Older records may contain only first/latest responses; missing intermediate revisions are not reconstructed.
Each field allows 2,000 characters. Keep a specific claim, state where it applies, cite the facts you use, and explain what still needs checking. Blank fields are allowed and are not evidence of mastery.
Prove if g(y)=Σx≤y f(x), then f(y)=Σx≤y μ(x, y)g(x).
Structure your argument and preserve revisions
These optional planning fields save locally. Recording your response saves an immutable copy of the argument and these fields. Older records may contain only first/latest responses; missing intermediate revisions are not reconstructed.
Each field allows 2,000 characters. Keep a specific claim, state where it applies, cite the facts you use, and explain what still needs checking. Blank fields are allowed and are not evidence of mastery.
Prove μ(A, B)=(−1)^(size (B−A)) in a Boolean interval.
Structure your argument and preserve revisions
These optional planning fields save locally. Recording your response saves an immutable copy of the argument and these fields. Older records may contain only first/latest responses; missing intermediate revisions are not reconstructed.
Each field allows 2,000 characters. Keep a specific claim, state where it applies, cite the facts you use, and explain what still needs checking. Blank fields are allowed and are not evidence of mastery.
Relate the divisor-poset μ(a, b) to arithmetic Möbius values.
Structure your argument and preserve revisions
These optional planning fields save locally. Recording your response saves an immutable copy of the argument and these fields. Older records may contain only first/latest responses; missing intermediate revisions are not reconstructed.
Each field allows 2,000 characters. Keep a specific claim, state where it applies, cite the facts you use, and explain what still needs checking. Blank fields are allowed and are not evidence of mastery.
Reflect and continue
Before marking this lesson complete, explain one key idea without the worked solution and identify a condition that matters. Marking complete records participation, not independent mastery.
