The Harbor City Ledger
Why QM Needs Inner-Product Geometry
A campus seminar turns abstract vectors into a practical language for amplitudes, probabilities and what measurement does — and does not — do.
SCIENCE & IDEAS
HARBOR CITY — Jan. 3, 2026
By Lena Park

A crowded lecture hall at Harbor City University fell unusually quiet Thursday as a visiting physicist drew a single arrow on the board and told students it was the beginning of almost everything they would calculate in quantum mechanics.
Professor Mara Ellison, invited from North Coast Institute, framed the day’s seminar around what she called “the chain that keeps quantum theory from turning into wordplay”: vector space to inner product to norm and orthogonality to projection.
The talk, hosted by the Department of Physics and a student-run “Math for Physicists” group, was billed as a refresher. It became, for many, a reconsideration of why quantum mechanics leans so heavily on geometric language.
“People say the state is a vector and then move on,” Ellison said in an interview afterward, packing chalk into a coat pocket. “But if you don’t commit to the geometry — specifically the inner product — you can’t tell the difference between an overlap and a probability.”
From ‘a place to add vectors’ to a way to compare them
Ellison began with what she called the “bare minimum”: a vector space.
On the board, she wrote a list of allowed moves — add two vectors, multiply by a scalar — and emphasized that none of those operations, by themselves, tells you whether two vectors are “close,” “perpendicular,” or “the same size.”
That gap is where the inner product enters.
“In quantum mechanics, you don’t just need a place for states to live,” she told the audience. “You need a rule for comparing states that respects complex numbers and still behaves like geometry.”
Students watched as she introduced the bracket notation common in the field, treating it less like symbolism and more like a measuring tool. The inner product, she said, takes two state vectors and returns a complex number — a number that carries phase information but can also be converted into something directly measurable.
Norm and orthogonality: size and ‘no overlap’
With the inner product in hand, Ellison moved to the norm — the notion of a vector’s length.
The norm, she said, is built from the inner product of a state with itself, and it is what allows physicists to enforce a basic operational demand: that the state be normalized.
“A normalized state is a promise,” she said. “It’s the promise that when you ask all possible outcomes the same question, the total weight adds up the way a probability should.”
She then connected that to orthogonality. Two states are orthogonal, she told students, when their inner product is zero. In that case, “there is no overlap,” meaning no component of one lies along the other.
A sophomore, Daniel Rios, said after the seminar that the language helped.
“I used to memorize rules about orthogonal states without understanding why they mattered,” he said. “Hearing it as ‘zero overlap’ makes it feel like geometry again.”
Projection: selecting a component, not rewriting the state’s past
The final link in Ellison’s chain was projection.
On the board she drew a state vector and a second vector labeled as a target state. Then she described projection as the operation that extracts the component of one vector along another.
“This is where measurement starts to look like geometry,” she told the room.
In the seminar’s most repeated phrase, she distinguished the “component” — the complex overlap — from the “weight,” the magnitude-squared that behaves like probability.
“Amplitude is a component; it can be complex,” she said, tapping the chalk beneath an inner product expression. “The weight is what survives when you ask, ‘How often would this outcome occur?’ It’s the magnitude-squared.”
She did not linger on philosophical claims about what measurement means, but she did insist on a limited, practical interpretation.
“A projection doesn’t inject energy into a system,” she said, responding to a question from the back row. “It selects a component relative to a question you’re asking. Confusing selection with physical forcing is a shortcut to bad intuition.”
Graduate teaching assistant Priya Nandakumar, who runs weekly problem sessions, said that confusion is common.
“Students will treat projection like a mysterious physical kick,” she said. “But in calculations, it’s a way of isolating the part of the state aligned with an outcome.”
Amplitude vs. weight: the distinction that keeps answers honest
Ellison returned several times to what she called the “two-level” nature of quantum calculations: first an overlap, then a weight.
A student can compute an overlap — a component or amplitude — and still not have a number that can be compared to experimental frequencies, she said. The overlap can carry a phase, which matters when multiple paths or states combine.
Only after taking the magnitude-squared does the expression become a weight, the quantity that, when normalized and tallied over outcomes, matches what experimenters count.
“That is why the inner product is not cosmetic,” she said. “It is the bookkeeping rule that turns geometry into statistics without losing the phase information you need for interference.”
A short checklist from the day’s notes
At the end of the seminar, organizers handed out a one-page summary with common mistakes. Several students lingered to compare their own habits to the list.
- Complex conjugation placement: In inner products involving complex vectors, putting the conjugate on the wrong side can flip phases and quietly corrupt an answer.
- Orthogonality vs. independence: Treating “orthogonal” as merely “different” leads to incorrect conclusions about overlap; orthogonality is a specific zero-inner-product condition, not a vague separation.
- Forgetting normalization: Using an unnormalized state without accounting for its norm can turn weights into numbers that do not add up consistently.
- Projection as energy change: Thinking projection “changes energy” rather than selecting a component can mislead intuition; in calculations, projection is a selection rule tied to the measurement basis.
Three prompts students were asked to answer for themselves
Before dismissing the room, Ellison asked students to write responses — in plain language — to three questions.
- In your own words, what extra power does an inner product add to a vector space when you are describing quantum states?
- How do you explain the difference between an overlap/component (amplitude) and a weight (magnitude-squared) without using equations?
- When you “project” a state onto a basis vector, what are you doing operationally, and what are you not claiming is happening physically?