Home / Courses / GATE Physics / Practice / Linear vector spaces and Dirac notation

Linear vector spaces and Dirac notation

Mathematical methods · 77 questions and templates for GATE Physics, in order from warm-up to top rank. Check each answer, then read the full solution.

New to this topic? Read the lesson first: Linear vector spaces and Dirac notation.

Loading questions… (they need JavaScript; nothing you type is sent anywhere)

Written questions with solutions

The same questions as a list, so you can print them or read without the practice box. Templates (fresh numbers) appear only in the practice box above.

  1. MCQWarm-up

    Which of these sets, with the usual addition and multiplication by real numbers, is a real vector space?
    1. All real 2×2 matrices with trace zero
    2. All polynomials of degree exactly 3
    3. All vectors (x,y,z) in ℝ3 with x+y+z=1
    4. All real 2×2 matrices with determinant zero
    Answer and solution

    Answer: A

    A subset is a vector space (a subspace) when it contains the zero vector and is closed under addition and scalar multiplication. Trace is linear: tr(aA+bB)=atrA+btrB, so traceless matrices stay traceless, and the zero matrix is traceless. Degree exactly 3 fails: x3+(−x3+x)=x has degree 1, and the zero polynomial is missing. The plane x+y+z=1 does not contain the zero vector. Determinant is not linear: diag(1,0)+diag(0,1)=I has determinant 1.

    Why people get this wrong: Check the zero vector first: a set that misses it can never be a vector space.

  2. NATWarm-up

    The dimension of the real vector space of all real symmetric 3×3 matrices is ______.
    Answer and solution

    Answer: 6

    A symmetric matrix is fixed by its diagonal (3 entries) and the entries above the diagonal (3 entries); the ones below are copies. A basis is the 3 matrices with a single 1 on the diagonal and the 3 matrices with 1 in positions (i,j) and (j,i). So the dimension is 3+3=6, or in general n(n+1)/2.

    Why people get this wrong: Counting all 9 entries ignores the constraint Aij=Aji.

  3. MCQWarm-up

    In ℂ2 with ⟨u|v⟩=∑kuk∗vk, the inner product of |u⟩=(1,i)T with |v⟩=(i,1)T is
    1. 2i
    2. 2
    3. 0
    4. −2i
    Answer and solution

    Answer: C

    The bra conjugates the components of |u⟩: ⟨u|=(1,−i). So ⟨u|v⟩=1⋅i+(−i)⋅1=0. The two vectors are orthogonal.

    Why people get this wrong: Forgetting to conjugate gives 1⋅i+i⋅1=2i.

  4. MCQWarm-up

    The norm of the vector (1+i, 2, −i)T in ℂ3 (standard inner product) is
    1. 5
    2. 7
    3. 3
    4. 7
    Answer and solution

    Answer: D

    ‖v‖2=∑|vk|2=|1+i|2+|2|2+|−i|2=2+4+1=7, so ‖v‖=7.

    Why people get this wrong: Squaring the components instead of taking |vk|2: (1+i)2+4+(−i)2=3+2i, which is not even real.

  5. NATWarm-up

    The set of all 2×2 complex Hermitian matrices is a vector space over the real numbers. Its dimension is ______.
    Answer and solution

    Answer: 4

    A Hermitian matrix (ab−icb+icd) has real a,b,c,d: four real parameters. A basis is {I,σx,σy,σz}, so the dimension is 4. (It is not a complex vector space: i times a Hermitian matrix is anti-Hermitian.)

    Why people get this wrong: Answering 8 (the real dimension of all complex 2×2 matrices) or 3 (forgetting the identity).

  6. MCQWarm-up

    The vectors (1,1,0), (0,1,1) and (1,0,1) in ℝ3
    1. are linearly dependent
    2. are linearly independent and form a basis of ℝ3
    3. are independent but span only a plane
    4. span a line
    Answer and solution

    Answer: B

    Put them as rows of a matrix: det⁡(110011101)=1(1−0)−1(0−1)+0=2≠0. Non-zero determinant means independent, and three independent vectors in a 3-dimensional space always form a basis.

    Why people get this wrong: Three independent vectors cannot "span only a plane": independence plus the right count already means a basis.

  7. MCQWarm-up

    The real vector space of all polynomials in x of degree at most 4 (with real coefficients) has dimension
    1. 4
    2. 5
    3. 6
    4. infinite
    Answer and solution

    Answer: B

    A basis is {1,x,x2,x3,x4}: every such polynomial is a unique combination of these five, and none of them is a combination of the others. So the dimension is 5.

    Why people get this wrong: Forgetting the constant term 1 gives 4.

  8. MSQStandard

    Which of these maps, acting on differentiable functions f(x), are linear operators?
    1. f↦dfdx
    2. f↦f2
    3. f↦xf(x)
    4. f↦f(x)+1
    Answer and solution

    Answer: A, C

    Linear means Oˆ(af+bg)=aOˆf+bOˆg. Differentiation and multiplication by x pass this test. Squaring gives (f+g)2≠f2+g2. Adding 1 fails because it does not send the zero function to zero (Oˆ(0)=1).

    Why people get this wrong: Adding a constant looks harmless, but a linear operator must map zero to zero.

  9. NATStandard

    The ket |ψ⟩=N(3|1⟩+4i|2⟩), with {|1⟩,|2⟩} orthonormal, is normalised for a real positive N equal to ______.
    Answer and solution

    Answer: 0.2 (±0.001)

    ⟨ψ|ψ⟩=N2(|3|2+|4i|2)=25N2=1, so N=1/5=0.2.

    Why people get this wrong: Using (4i)2=−16 gives 9−16<0; the norm uses |4i|2=16.

  10. MCQStandard

    With |a⟩=(1,0)T and |b⟩=(i,1)T, the outer product |a⟩⟨b| is the matrix
    1. (−i100)
    2. (i100)
    3. (−i010)
    4. (i010)
    Answer and solution

    Answer: A

    ⟨b| is the conjugate transpose of |b⟩: the row (−i, 1). A column times a row: (10)(−i, 1)=(−i100).

    Why people get this wrong: Forgetting the conjugate gives i in the corner; writing the row as a column gives the transposed answers.

  11. NATStandard

    Gram-Schmidt is applied to v→1=(1,1,0) and v→2=(1,0,1): u→2=v→2−v→2⋅v→1v→1⋅v→1v→1. The value of |u→2|2 is ______.
    Answer and solution

    Answer: 1.5 (±0.01)

    v→2⋅v→1=1 and v→1⋅v→1=2, so u→2=(1,0,1)−12(1,1,0)=(12,−12,1) and |u→2|2=14+14+1=1.5. Check: |v→2|2−(v→2⋅v→1)2/|v→1|2=2−12=1.5 (Pythagoras).

    Why people get this wrong: Dividing by |v→1| instead of |v→1|2 in the projection.

  12. MCQStandard

    The Cauchy-Schwarz inequality |⟨u|v⟩|2≤⟨u|u⟩⟨v|v⟩ becomes an equality exactly when
    1. |u⟩ and |v⟩ are orthogonal
    2. |u⟩ and |v⟩ have the same norm
    3. one of the vectors is a (complex) multiple of the other
    4. both vectors have real components
    Answer and solution

    Answer: C

    Write |v⟩=c|u⟩+|w⟩ with ⟨u|w⟩=0. Then |⟨u|v⟩|2=|c|2⟨u|u⟩2 while ⟨u|u⟩⟨v|v⟩=⟨u|u⟩(|c|2⟨u|u⟩+⟨w|w⟩). They are equal only when ⟨w|w⟩=0, that is |w⟩=0: the vectors are parallel.

    Why people get this wrong: Orthogonal vectors give the opposite extreme: the left side is zero.

  13. MCQStandard

    For an orthonormal basis {|n⟩} of a finite-dimensional space, the operator ∑n|n⟩⟨n| equals
    1. the zero operator
    2. the number equal to the dimension
    3. the projector onto |1⟩ only
    4. the identity operator
    Answer and solution

    Answer: D

    Any vector expands as |ψ⟩=∑ncn|n⟩ with cn=⟨n|ψ⟩, so |ψ⟩=∑n|n⟩⟨n|ψ⟩ for every |ψ⟩. That says ∑n|n⟩⟨n|=1ˆ: the completeness relation. (Its trace is the dimension, which is where the second option comes from.)

    Why people get this wrong: ∑n⟨n|n⟩ (a number) is not the same as ∑n|n⟩⟨n| (an operator).

  14. NATStandard

    For the (unnormalised) ket |ψ⟩=(1, 2i, 2)T, the trace of the operator |ψ⟩⟨ψ| is ______.
    Answer and solution

    Answer: 9

    tr(|ψ⟩⟨ψ|)=∑kψkψk∗=⟨ψ|ψ⟩=1+4+4=9. In general tr(|a⟩⟨b|)=⟨b|a⟩.

    Why people get this wrong: Summing ψk2 without the conjugate gives 1−4+4=1.

  15. MCQStandard

    In the orthonormal basis |e1⟩=(1,1)/2, |e2⟩=(1,−1)/2, the component of |v⟩=(3,1) along |e2⟩ is
    1. 22
    2. 2
    3. 1
    4. 2
    Answer and solution

    Answer: B

    In an orthonormal basis the component is the inner product: ⟨e2|v⟩=(3−1)/2=2/2=2. (Along |e1⟩: 4/2=22; check (2)2+(22)2=10=|v|2.)

    Why people get this wrong: Forgetting the 1/2 gives 2; taking the e1 component gives 22.

  16. MSQStandard

    For u→,v→∈ℝ2, which of these define a valid inner product ⟨u→,v→⟩?
    1. u1v1+2u2v2
    2. u1v1−u2v2
    3. u1v2+u2v1
    4. 2u1v1+u1v2+u2v1+u2v2
    Answer and solution

    Answer: A, D

    Each is u→TGv→ for a symmetric G; it is an inner product when G is positive definite. A: G=diag(1,2), positive. B: diag(1,−1) gives ⟨v→,v→⟩=−1 for (0,1). C: G=(0110) gives −2 for (1,−1). D: G=(2111) has trace 3 and determinant 1, so both eigenvalues are positive: valid.

    Why people get this wrong: Symmetry and linearity are easy; positivity (⟨v,v⟩>0 for v≠0) is where B and C fail.

  17. NATStandard

    The dimension of the subspace of ℝ4 spanned by (1,0,1,0), (0,1,0,1), (1,1,1,1) and (1,−1,1,−1) is ______.
    Answer and solution

    Answer: 2

    Call them a,b,c,d. Then c=a+b and d=a−b, while a and b are independent. So the span is the plane spanned by a and b: dimension 2.

    Why people get this wrong: Four vectors do not mean four dimensions; look for combinations first.

  18. MCQStandard

    Consider smooth real functions f on [0,L] with the usual addition and scaling. Which set is NOT a vector space?
    1. All f with f(0)=f(L)=0
    2. All f with f′(0)=2f(0)
    3. All f with f(0)=1
    4. All f with f(0)=f(L)
    Answer and solution

    Answer: C

    A subset is a vector space exactly when it contains the zero function and is closed under addition and scaling. Linear, homogeneous conditions (the first, second and fourth) pass: if f and g satisfy them, so does af+bg. The condition f(0)=1 fails: the zero function is not in the set, and the sum of two such functions has f(0)+g(0)=2. This is why boundary conditions in eigenvalue problems are kept homogeneous.

    Why people get this wrong: A Robin condition like f′(0)=2f(0) looks complicated but is still linear and homogeneous.

  19. NATStandard

    With the inner product ⟨f,g⟩=∫01f(x)g(x)dx, the angle between the functions f=1 and g=x is ______ degrees.
    Answer and solution

    Answer: 30 (±0.5)

    ⟨1,x⟩=∫01xdx=12, ‖1‖2=1, ‖x‖2=∫01x2dx=13. So cos⁡θ=1/21⋅1/3=32 and θ=30∘.

    Why people get this wrong: Forgetting to take the square root of ‖x‖2 gives cos⁡θ=3/2, which is impossible.

  20. MCQExam level

    With {|1⟩,|2⟩} orthonormal and Aˆ=|1⟩⟨2|, the operator Aˆ†Aˆ equals
    1. |1⟩⟨1|
    2. |2⟩⟨2|
    3. 0
    4. 1ˆ
    Answer and solution

    Answer: B

    Aˆ†=(|1⟩⟨2|)†=|2⟩⟨1|. So Aˆ†Aˆ=|2⟩⟨1|1⟩⟨2|=|2⟩⟨2|, using ⟨1|1⟩=1.

    Why people get this wrong: |1⟩⟨1| is AˆAˆ†; the order matters.

  21. NATExam level

    An operator is Aˆ=|1⟩⟨1|+2|2⟩⟨2|+3|3⟩⟨3| in an orthonormal basis. For |ψ⟩=13(|1⟩+2i|2⟩+2|3⟩), the value of ⟨ψ|Aˆ|ψ⟩ is ______ (to two decimal places).
    Answer and solution

    Answer: 2.33 (±0.01)

    First check the norm: (1+4+4)/9=1. Then ⟨Aˆ⟩=∑nan|cn|2=1⋅19+2⋅49+3⋅49=219≈2.33.

    Why people get this wrong: The phase i does not matter: only |cn|2 enters.

  22. MCQExam level

    Let |±⟩=(|0⟩±|1⟩)/2, where σz|0⟩=|0⟩ and σz|1⟩=−|1⟩. The matrix of σz in the basis {|+⟩,|−⟩} is
    1. (100−1)
    2. (0−ii0)
    3. (0110)
    4. 12(111−1)
    Answer and solution

    Answer: C

    σz|+⟩=(|0⟩−|1⟩)/2=|−⟩ and σz|−⟩=|+⟩. So ⟨+|σz|+⟩=0, ⟨−|σz|+⟩=1, ⟨+|σz|−⟩=1, ⟨−|σz|−⟩=0. The matrix is (0110): in the x basis, σz looks like σx.

    Why people get this wrong: An operator keeps its eigenvalues in every basis, but its matrix changes; it is diagonal only in its own eigenbasis.

  23. MSQExam level

    For a Hermitian operator on a finite-dimensional complex inner-product space, which statements are always true?
    1. Its eigenvalues are real
    2. Eigenvectors belonging to different eigenvalues are orthogonal
    3. Its eigenvectors can be chosen to form an orthonormal basis
    4. It is invertible
    Answer and solution

    Answer: A, B, C

    Real eigenvalues and orthogonality for distinct eigenvalues follow from ⟨a|Hˆb⟩=⟨Hˆa|b⟩. The spectral theorem adds completeness: an orthonormal eigenbasis exists (degenerate eigenspaces are orthonormalised by Gram-Schmidt). Invertibility fails when 0 is an eigenvalue, for example |1⟩⟨1|.

    Why people get this wrong: Hermitian does not mean invertible: a projector is Hermitian with eigenvalue 0.

  24. NATExam level

    For |a⟩=(1, i, 0)T and |b⟩=(0, 1, i)T, define the angle θ between them by cos⁡θ=|⟨a|b⟩|/(‖a‖‖b‖). Then θ in degrees is ______.
    Answer and solution

    Answer: 60 (±0.5)

    ⟨a|=(1,−i,0), so ⟨a|b⟩=0+(−i)(1)+0=−i and |⟨a|b⟩|=1. Both norms are 2. So cos⁡θ=1/2 and θ=60∘.

    Why people get this wrong: Without the conjugate you get i, the same modulus here, but the habit costs marks elsewhere.

  25. MCQExam level

    With the inner product ⟨f,g⟩=∫−11f(x)g(x)dx, which pair of functions is orthogonal?
    1. 1 and x2
    2. x and x3
    3. 1 and x4
    4. x and x2
    Answer and solution

    Answer: D

    x⋅x2=x3 is odd, so its integral over the symmetric interval vanishes. The other products are even and positive: ∫x2=23, ∫x4=25 (twice: for 1⋅x4 and x⋅x3).

    Why people get this wrong: Different powers are not automatically orthogonal; parity decides here.

  26. NATExam level

    Gram-Schmidt on {1,x,x2} with ⟨f,g⟩=∫−11fgdx gives a third orthogonal polynomial of the form x2−c. The constant c is ______ (to three decimal places).
    Answer and solution

    Answer: 0.333 (±0.002)

    Subtract the projections of x2 on the first two: ⟨x2,x⟩=0 (odd), and ⟨x2,1⟩/⟨1,1⟩=(2/3)/2=1/3. So the polynomial is x2−13, which is proportional to the Legendre polynomial P2=12(3x2−1).

    Why people get this wrong: Forgetting to divide by ⟨1,1⟩=2 gives c=2/3.

  27. MCQExam level

    With {|1⟩,|2⟩} orthonormal, Aˆ=|1⟩⟨2| and Bˆ=|2⟩⟨1|, the commutator [Aˆ,Bˆ] is
    1. |1⟩⟨1|−|2⟩⟨2|
    2. 0
    3. 1ˆ
    4. |2⟩⟨2|−|1⟩⟨1|
    Answer and solution

    Answer: A

    AˆBˆ=|1⟩⟨2|2⟩⟨1|=|1⟩⟨1| and BˆAˆ=|2⟩⟨1|1⟩⟨2|=|2⟩⟨2|. So [Aˆ,Bˆ]=|1⟩⟨1|−|2⟩⟨2|, which is σz (these are the raising and lowering operators σ±).

    Why people get this wrong: Getting the sign backwards by computing BˆAˆ−AˆBˆ.

  28. MSQExam level

    The complex 2×2 matrices form an inner-product space with ⟨A,B⟩=tr(A†B). Which statements are true?
    1. I,σx,σy,σz are mutually orthogonal
    2. Each Pauli matrix has norm 2
    3. The space has complex dimension 4
    4. σx and iσx are orthogonal
    Answer and solution

    Answer: A, B, C

    The Pauli matrices are traceless and Hermitian with σjσk=δjkI+iϵjklσl, so tr(σjσk)=2δjk and tr(Iσj)=0: orthogonal, each of norm 2 (as is I). Four orthogonal non-zero matrices in the 4-dimensional space of 2×2 matrices form a basis. But ⟨σx,iσx⟩=itr(σx2)=2i≠0.

    Why people get this wrong: Over the complex numbers, iσx is a multiple of σx: parallel, not orthogonal.

  29. NATExam level

    Write M=(31−i1+i−1)=a0I+a→⋅σ→. The larger eigenvalue of M, a0+|a→|, is ______ (to two decimal places).
    Answer and solution

    Answer: 3.45 (±0.01)

    Use aμ=12tr(Mσμ): a0=12(3−1)=1, az=12(3+1)=2, ax=ReM21=1, ay=ImM21=1. Since (a→⋅σ→)2=|a→|2I, the eigenvalues are a0±|a→|=1±6. Larger: 1+2.449=3.45. Check: trace 2 and determinant −3−2=−5 give λ2−2λ−5=0, the same roots.

    Why people get this wrong: Reading ay from M12=1−i with the wrong sign does not change |a→| here, but it does change the eigenvectors.

  30. MCQExam level

    For a complex number c, the bra corresponding to the ket c|ψ⟩ is
    1. c⟨ψ|
    2. c∗⟨ψ|
    3. |c|⟨ψ|
    4. ⟨ψ|/c
    Answer and solution

    Answer: B

    The map from kets to bras is antilinear: ⟨ϕ|(cψ)⟩ must equal c⟨ϕ|ψ⟩, which forces the bra of c|ψ⟩ to be c∗⟨ψ| (so that ⟨cψ|ϕ⟩=c∗⟨ψ|ϕ⟩).

    Why people get this wrong: Treating the bra as a plain transpose and forgetting the complex conjugate.

← All topics · Timed mock test →