Linear combinations and span: deciding membership using finite coefficient lists as witnesses

On this page
  1. Goal and scope
  2. Givens and symbols
  3. Rules and sources
  4. Main derivation
  5. Substeps and lemmas
  6. Reverse audits
  7. Error controls and negative controls
  8. Result and limitations
  9. Verification and source binding

← Guide index · Original MKO

derivation-mko-linear-combination-span-01 · production-20261019-adopted-96

Goal and scope

(1) In ℝ², with v₁=(1,2) and v₂=(3,−1), decide whether b=(7,0) lies in span{v₁,v₂}: first prove that the span of two vectors is exactly {c₁v₁+c₂v₂}, then solve for the coefficients, prove they are unique, and rebuild b from the witness. (2) In ℝ³, with u₁=(1,0,1) and u₂=(0,1,1), prove that t=(2,3,4) is not in span{u₁,u₂} while t'=(2,3,5) is: having three coordinates does not decide membership.

Audits only the explicitly given finite generator families and candidate witnesses (MKO assumption a2), with integer-coordinate exercise vectors and worked membership witnesses whose coefficients happen to be integers (MKO assumption a1); the general span definition and two-generator lemma allow all real coefficients. The span is defined using finite-length linear combinations. An infinite-series expression is not such a representation; if its sum exists, membership is decided by whether that sum has a finite representation. Linear independence, bases and dimension are not treated; the canonical MKO and its English edition are not changed.

Givens and symbols

  • vector_space_axioms
  • eq_congruence
  • mul_inverse_nonzero
  • set_extensionality
  • numeric_evaluation
  • G1 — In ℝⁿ, addition and scalar multiplication act componentwise; v₁=(1,2), v₂=(3,−1) and b=(7,0) all lie in ℝ².
  • G2 — u₁=(1,0,1), u₂=(0,1,1), t=(2,3,4) and t'=(2,3,5) all lie in ℝ³.
  • G3 — Each question audits only the explicitly given finite family and candidate witness; conclusions do not extend to other families (MKO assumption a2).
  • v_1,\ v_2 — the generator family in ℝ²
  • u_1,\ u_2 — the generator family in ℝ³
  • c_1,\ c_2 — the real coefficients of a linear combination
  • \operatorname{span}\{w_1,w_2\} — the set of all finite linear combinations
  • (c_1,c_2) — a membership witness

Rules and sources

  • vector_space_axioms · prerequisite — Vector space axioms (over ℝ or another field F): vectors form a commutative group under addition; scalar multiplication satisfies a(u+v)=au+av, (a+b)u=au+bu, (ab)u=a(bu) and 1u=u.
  • eq_congruence · prerequisite — Congruence of equality: if u=v, applying the same defined operation f to both sides gives f(u)=f(v). Operating on one side only is not an instance of this rule.
  • mul_inverse_nonzero · prerequisite — Inverse of a nonzero number: in ℚ, ℝ or ℂ, if u≠0 there is 1/u with u·(1/u)=1, and v/u means v·(1/u). 0 has no inverse, so v/0 is undefined.
  • set_extensionality · prerequisite — Set equality: A=B if and only if A⊆B and B⊆A, i.e. they have the same elements.
  • numeric_evaluation · prerequisite — Exact evaluation of an explicit integer or rational expression; independently recomputable.
  • def_intro · prerequisite — Introducing a definition: a statement, expression or object may be given a name, after which the name and what it stands for are interchangeable.
  • finite_sum_rules · prerequisite — Finite sums: Σ_{i=1}^{0} a_i=0 and Σ_{i=1}^{n+1} a_i=Σ_{i=1}^{n} a_i + a_{n+1}; Σ(a_i+b_i)=Σa_i+Σb_i; Σ c·a_i=c·Σa_i.
  • def_unfold · prerequisite — Unfolding a definition: replacing a defined term by its definition (or folding it back) does not change whether a statement is true.
  • set_builder · prerequisite — Set-builder notation: x∈{y∈A ∣ P(y)} if and only if x∈A and P(x).
  • mul_comm · prerequisite — Commutativity of multiplication: uv=vu (in commutative number systems such as ℤ, ℚ, ℝ, ℂ).
  • neg_mul · prerequisite — Signs and multiplication: (−u)v=−(uv) and (−u)(−v)=uv.
  • add_inverse · prerequisite — Additive inverse: every u has −u with u+(−u)=0; u−v means u+(−v).
  • add_zero · prerequisite — Additive identity: u+0=u.
  • eq_replace · prerequisite — Replacement of equals: an expression inside a statement may be replaced by an equal expression without changing whether the statement is true.
  • distributive · prerequisite — Distributivity: u(v+w)=uv+uw; with commutativity also (v+w)u=vu+wu.
  • logic_chain · prerequisite — Chaining: from P⇒Q and Q⇒R infer P⇒R; likewise for ⇔.
  • exists_intro · prerequisite — ∃-introduction: if some a∈D satisfies P(a), then ∃x∈D, P(x).
  • contradiction · prerequisite — Proof by contradiction: if assuming ¬P leads to a contradiction (a statement and its negation both hold), then P holds.
  • and_intro_elim · prerequisite — Conjunction: from P and Q infer P∧Q; from P∧Q infer P and infer Q.
  • mko_span_def · cited — For a vector family in a real vector space, a finite linear combination is a finite sum weighted by real coefficients. Its span is the set of all finite-length linear combinations and does not include undefined infinite series. · mko-linear-combination-span
  • coord_ops · prerequisite — In ℝⁿ of fixed dimension n, vector addition and scalar multiplication act componentwise; vectors of different dimensions cannot be added.
  • tuple_eq · prerequisite — Two ordered tuples of the same length are equal if and only if all corresponding components are equal.
  • zero_scalar · prerequisite — In a real vector space 0·w=0 (the zero vector): 0w=(0+0)w=0w+0w, and adding −(0w) to both sides gives it; the empty finite sum is defined to be the zero vector.
  • det2_criterion · prerequisite — A 2×2 linear system px+qy=e, rx+sy=f with ps−qr≠0 has exactly one solution (Cramer's rule).
  • L_span_two · lemma — For a family (w₁,w₂) in a real vector space: span{w₁,w₂}={c₁w₁+c₂w₂ ∣ c₁,c₂∈ℝ}.

Main derivation

S01 · definition

Before

\text{family } (w_1,w_2)

After

\operatorname{span}\{w_1,w_2\}=\{\textstyle\sum_{k=1}^{n}t_kx_k\mid n\in\mathbb{N},\ t_k\in\mathbb{R},\ x_k\in\{w_1,w_2\}\}

Conditions: none

Rule: mko_span_def · Uses: mko_span_def

Leaf basis: mko_span_def + def_intro

By the MKO, a finite linear combination is a finite sum weighted by real coefficients, and the span is the set of all finite-length combinations. The length n may be any natural number (n=0 gives the empty sum, the zero vector), but each combination has finitely many terms; an infinite-series expression is not the finite representation used by this definition, but its sum, when defined, may have a finite representation and lie in the span.

S02 · implies

Before

v_1,v_2,b

After

v_1,v_2,b\in\mathbb{R}^2;\ \text{witness }(c_1,c_2)\in\mathbb{R}^2

Conditions: none

Rule: coord_ops · Uses: G1, coord_ops

Leaf basis: coord_ops

All three vectors have two components and lie in the same ℝ², so weighted sums are defined (MKO paragraph 1). The family has two vectors, so a witness has exactly two coefficients.

S03 · equivalent

Before

\operatorname{span}\{w_1,w_2\}

After

\{c_1w_1+c_2w_2\mid c_1,c_2\in\mathbb{R}\}

Conditions: none

Rule: L_span_two · Uses: S01, vector_space_axioms

Leaf basis: set_extensionality + vector_space_axioms + zero_scalar

The definition allows any length and repeated use of the generators; Lemma L_span_two shows they all collapse to a form with exactly two coefficients. The proof is in S03.1 (⊆) and S03.2 (⊇), and set extensionality gives the equality.

S04 · equivalent

Before

b\in\operatorname{span}\{v_1,v_2\}

After

\exists c_1,c_2\in\mathbb{R}:\ c_1(1,2)+c_2(3,-1)=(7,0)

Conditions: none

Rule: set_builder · Uses: S03, S02

Leaf basis: set_builder + L_span_two

By S03, membership is equivalent to the existence of two real coefficients whose weighted sum is b. Such a pair (c₁,c₂) is what the MKO calls a membership witness.

S05 · equivalent

Before

c_1(1,2)+c_2(3,-1)

After

(c_1+3c_2,\ 2c_1-c_2)

Conditions: none

Rule: coord_ops · Uses: G1, coord_ops

Leaf basis: coord_ops + mul_comm + neg_mul

Componentwise: first component c₁·1+c₂·3=c₁+3c₂, second component c₁·2+c₂·(−1)=2c₁−c₂.

S06 · equivalent

Before

(c_1+3c_2,\ 2c_1-c_2)=(7,0)

After

c_1+3c_2=7\ \land\ 2c_1-c_2=0

Conditions: none

Rule: tuple_eq · Uses: S04, S05, tuple_eq

Leaf basis: tuple_eq

Ordered pairs are equal exactly when both components agree, so the vector equation becomes two real equations, which one pair (c₁,c₂) must satisfy simultaneously.

S07 · equivalent

Before

2c_1-c_2=0

After

c_2=2c_1

Conditions: none

Rule: eq_congruence · Uses: S06

Leaf basis: eq_congruence + add_inverse + add_zero

Add c₂ to both sides to get 2c₁=c₂. The backward move subtracts c₂ from both sides.

S08 · equivalent

Before

c_1+3c_2=7,\ c_2=2c_1

After

(7c_1=7\ \land\ c_2=2c_1)\iff(c_1=1\ \land\ c_2=2c_1)

Conditions: 7\neq 0

Rule: mul_inverse_nonzero · Uses: S06, S07

Leaf basis: eq_replace + distributive + mul_inverse_nonzero + numeric_evaluation

Substitute c₂=2c₁ into the first equation: c₁+6c₁=7c₁=7. Since 7≠0, multiplying both sides by 1/7 gives c₁=1. Given c₂=2c₁, the substitution is an equivalence.

S09 · equivalent

Before

c_1=1,\ c_2=2c_1

After

(c_1,c_2)=(1,2)\ \text{(unique)}

Conditions: none

Rule: eq_replace · Uses: S07, S08

Leaf basis: eq_replace + logic_chain + numeric_evaluation

c₂=2·1=2. Every step from S06 to S09 is an equivalence, so the system is equivalent to (c₁,c₂)=(1,2): a solution exists and it is unique.

S10 · implies

Before

1\cdot(1,2)+2\cdot(3,-1)

After

(1+6,\ 2-2)=(7,0)=b\ \Rightarrow\ b\in\operatorname{span}\{v_1,v_2\}

Conditions: none

Rule: exists_intro · Uses: S04, S09, G1

Leaf basis: exists_intro + coord_ops + numeric_evaluation

Rebuild the target from the witness (1,2): (1,2)+(6,−2)=(7,0)=b (MKO paragraph 2). By ∃-introduction and S04, b lies in span{v₁,v₂}. The rebuild is itself an independent check.

S11 · implies

Before

\det\begin{pmatrix}1&3\\2&-1\end{pmatrix}

After

1\cdot(-1)-3\cdot 2=-7\neq 0

Conditions: none

Rule: det2_criterion · Uses: S06, det2_criterion

Leaf basis: det2_criterion + numeric_evaluation

Cross-check: the coefficient determinant is −7≠0, so by Cramer's rule the system has exactly one solution, matching the unique solution found by substitution in S09 (MKO paragraph 3: two independent generators of ℝ² give unique coordinates via a nonzero determinant).

S12 · equivalent

Before

t\in\operatorname{span}\{u_1,u_2\}

After

\exists c_1,c_2:\ c_1=2\ \land\ c_2=3\ \land\ c_1+c_2=4

Conditions: none

Rule: tuple_eq · Uses: G2, S03, coord_ops, tuple_eq

Leaf basis: L_span_two + coord_ops + tuple_eq

Lemma L_span_two holds for any family of two vectors, so t∈span{u₁,u₂} is equivalent to the existence of c₁, c₂ with c₁(1,0,1)+c₂(0,1,1)=(c₁, c₂, c₁+c₂) equal to (2,3,4), which splits componentwise into three equations.

S13 · implies

Before

c_1=2\ \land\ c_2=3\ \land\ c_1+c_2=4

After

5=4\ \text{false};\ t\notin\operatorname{span}\{u_1,u_2\}

Conditions: none

Rule: contradiction · Uses: S12

Leaf basis: contradiction + eq_replace + numeric_evaluation

The first two equations force c₁=2 and c₂=3; substituting into the third gives 2+3=5=4, a contradiction. So no witness exists and t is not in span{u₁,u₂}, even though t has three coordinates like u₁ and u₂.

S14 · implies

Before

2u_1+3u_2

After

(2,0,2)+(0,3,3)=(2,3,5)=t'\ \Rightarrow\ t'\in\operatorname{span}\{u_1,u_2\}

Conditions: none

Rule: exists_intro · Uses: G2, S03

Leaf basis: exists_intro + coord_ops + numeric_evaluation

For t'=(2,3,5) the same first two equations give c₁=2 and c₂=3, and the third equation 2+3=5 holds. The witness (2,3) rebuilds t', so t'∈span{u₁,u₂}. t and t' differ only in the third component, yet their membership is opposite.

S15 · implies

Before

b,\ t,\ t'

After

b\in\operatorname{span}\{v_1,v_2\}\ \text{via unique }(1,2);\ t\notin,\ t'\in\operatorname{span}\{u_1,u_2\}

Conditions: none

Rule: and_intro_elim · Uses: S09, S10, S13, S14

Leaf basis: and_intro_elim

Combine the conclusions: b lies in span{v₁,v₂} via the unique witness (1,2); t is not in and t' is in span{u₁,u₂}. Each conclusion is decided by a finite witness or by a contradiction in the equations, never by dimension.

Substeps and lemmas

S03.1 · implies

Before

\sum_{k=1}^{n}t_kx_k,\ x_k\in\{w_1,w_2\}

After

\Big(\sum_{k\in K_1}t_k\Big)w_1+\Big(\sum_{k\in K_2}t_k\Big)w_2

Conditions: none

Rule: vector_space_axioms · Uses: S01, vector_space_axioms, zero_scalar

Leaf basis: vector_space_axioms + finite_sum_rules + zero_scalar

(⊆) Take any finite combination. For each term first choose a label j_k∈{1,2} with x_k=w_(j_k), then partition the indices by those chosen labels into K₁ and K₂. Every index belongs to exactly one group. Even if w₁=w₂, all labels may be 1 and K₂ empty; equal vector values do not themselves determine a partition. Reorder the terms by associativity and commutativity of vector addition, then merge coefficients of the same vector by (a+b)w=aw+bw, giving c₁w₁+c₂w₂ with c₁, c₂ the sums of the coefficients in each group. If a group is empty its coefficient sum is 0, and 0·w=0, so the form still holds.

S03.2 · implies

Before

c_1w_1+c_2w_2

After

\sum_{k=1}^{2}t_kx_k,\ t_1=c_1,\ x_1=w_1,\ t_2=c_2,\ x_2=w_2

Conditions: none

Rule: def_unfold · Uses: S03.1, S01

Leaf basis: def_unfold + set_extensionality

(⊇) c₁w₁+c₂w₂ is itself a finite linear combination of length 2, so it lies in the span. With both inclusions, set extensionality gives span{w₁,w₂}={c₁w₁+c₂w₂}. S03.1 and S03.2 together prove Lemma L_span_two.

Reverse audits

  • S15 · sufficient_goal_reduction · S09, S10, S13, S14, and_intro_elim — Goal G: the three membership decisions. Sufficient condition H: S09, S10, S13 and S14; H⇒G by conjunction introduction. H is derived forwards from the givens; G is never used as a premise.
  • S14 · premise_audit · S03, exists_intro, numeric_evaluation — Existence needs one witness, checked by rebuilding componentwise (N7, N8).
  • S13 · premise_audit · S12, contradiction, numeric_evaluation — The contradiction comes from one pair of coefficients having to satisfy all three equations; the first two leave no choice, so the refutation covers every possible pair.
  • S12 · inverse_equivalence · S03, coord_ops, tuple_eq — The lemma, the componentwise computation and tuple equality are all two-way.
  • S11 · premise_audit · det2_criterion, numeric_evaluation — Uniqueness is already proved by the chain of equivalences in S09; this step is a consistency check and the guide does not depend on Cramer's rule.
  • S10 · premise_audit · S04, exists_intro, numeric_evaluation — Existence needs one witness; uniqueness comes separately from S09.
  • S09 · inverse_equivalence · eq_replace, logic_chain — The chain of equivalences is joined by logic_chain; uniqueness comes from every link being two-way.
  • S08 · inverse_equivalence · eq_replace, mul_inverse_nonzero, distributive — Multiplying by 1/7 and by 7 undo each other, given 7≠0; substituting and substituting back are inverse moves.
  • S07 · inverse_equivalence · eq_congruence, add_inverse — Adding and subtracting c₂ undo each other, with no side condition.
  • S06 · inverse_equivalence · tuple_eq — The equality condition for ordered pairs is itself two-way.
  • S05 · inverse_equivalence · coord_ops, mul_comm — A componentwise identity rewrite (checked on a grid in N10, N11); two-way.
  • S04 · inverse_equivalence · S03, set_builder — Membership in set-builder notation is two-way.
  • S03.2 · premise_audit · S03.1, def_unfold, set_extensionality — This direction only needs to recognize a length-2 combination by the definition.
  • S03.1 · premise_audit · vector_space_axioms, zero_scalar, finite_sum_rules — Only finitely many reorderings and mergings are used, so this lemma treats finite combinations only. Convergence and the validity of reordering an infinite series require separate justification. This step proves ⊆ only.
  • S03 · inverse_equivalence · L_span_two, set_extensionality — Both inclusions are proved, so the sets are equal and the step is two-way.
  • S02 · premise_audit · G1, coord_ops — Only equal dimensions and the number of coefficients are checked; this does not show that b is in the span.
  • S01 · premise_audit · mko_span_def — This step only states the definition and derives no new fact; S03 reduces it to a form with exactly two coefficients.

Error controls and negative controls

An infinite series is not a finite linear combination

One mistake the MKO names is treating an infinite-series expression directly as a finite linear-combination representation. The expression Σ_{k=1}^{∞} 2^{−k}v₁ has infinitely many terms, but its sum here is v₁, which lies in the span and has the finite representation 1·v₁. Infinitely many terms do not imply that the sum is outside the span. S03.1 proves only finite reorderings and mergings; infinite series need separate convergence and reordering justification.

Dimension alone does not decide membership

Another mistake identified in the MKO is declaring a vector in the span by looking at its dimension. t and t' are both in ℝ³, the same dimension as u₁ and u₂, but t is not in span{u₁,u₂} while t' is. Membership must be decided by rebuilding a witness or by a contradiction in the equations.

Do not confuse coefficients with target components

The witness is the coefficient pair (c₁,c₂), not the components of the target b. Using (7,0) as coefficients gives 7(1,2)+0(3,−1)=(7,14)≠b. Both equations must also hold for one pair: c₁=7, c₂=0 satisfies the first equation but makes the second 14≠0.

  • t=(2,3,4) has the same dimension as u₁ and u₂, so t∈span{u₁,u₂} → reject — The equations demand c₁=2, c₂=3 and c₁+c₂=4, that is 5=4, so there is no solution.
  • Without rewriting its sum finitely, the displayed infinite-series expression Σ_{k=1}^{∞} 2^{−k}v₁ itself is a finite-length linear-combination representation allowed by the definition → reject — This expression has infinitely many terms and is not a finite linear-combination representation; nevertheless its sum here is v₁, which lies in the span.
  • (c₁,c₂)=(7,0) is a witness for b → reject — 7(1,2)+0(3,−1)=(7,14)≠(7,0).
  • Since c₁+3c₂=7 has a solution, b∈span{v₁,v₂} → reject — One pair must satisfy both equations; the conclusion happens to be true but the reason is invalid, since for example (7,0) satisfies only the first equation.

Result and limitations

The span of two vectors is exactly {c₁w₁+c₂w₂} (proved). b=(7,0)=1·v₁+2·v₂ with the unique witness (1,2) (the determinant −7≠0 is a cross-check), so b∈span{v₁,v₂}. For t=(2,3,4) the equations demand 5=4, so t∉span{u₁,u₂}, while t'=(2,3,5)=2u₁+3u₂∈span{u₁,u₂}. Having three coordinates does not decide membership.

  • Lemma L_span_two is written for families of two vectors only; any finite family works the same way (group the indices by the generator used), but this is not written out here. det2_criterion (Cramer's rule) is used only as a cross-check and is not proved here.
  • Linear independence, bases and dimension (the MKO's formal next) are not treated; the conclusions concern only the given generator families (MKO assumption a2).

Verification and source binding

11 declared checks for this guide; only the listed expressions are covered.

  • source commit: b28a5edbd46279b1db2a3f3f48bb3f447e18899f
  • adopted source SHA-256: 80267e70b6712e769dacf39909b666995788a4c88937ec1d58633f8a3b9b515f
  • original source SHA-256: 179ee4ab9d0ac1428d4346b01cc0e9b8f510b5eea8a68acc501365a2ddaddca5
  • Lean: NOT_RUN · independent AI: NOT_MEASURED

Same-source full JSON (formulas, ASTs, controls and revision log)