CONTROL
alistair
Oxford swells in the autumn.
However, the lecture theatre is noticeably underfilled this time.
Michaelmas term usually brings an excess of bodies—fresh faces packed shoulder to shoulder, latecomers lingering along the aisles, an almost feral eagerness to occupy space they haven’t yet earned.
Today, the room sits at roughly three-quarters capacity.
Not freshmen, of course. I don’t teach first-years. Juniors ask too many questions without understanding why the answers matter. Curiosity without structure is noise, and I have little tolerance for it.
My modules are reserved for final-year students—those who’ve already survived the attrition curve.
Those who understand that not all information is offered freely, not all problems are meant to be solved aloud.
And about the postgrads… they act and talk as if they know everything. I despise those who play-pretend.
I step inside, the door closing behind me with a soft, decisive click. The room stills almost immediately.
Conversations die mid-breath. Spines straighten. Screens tilt upward, attention snapping into place with near-instinctive precision. I don’t rush. I let the silence settle.
I cross to the lectern and place my iPad down neatly at its center. The podium bears the faint grooves of decades of lecturers before me—men who believed authority came from raised voices or exaggerated gestures. They were wrong. Authority comes from certainty.
I clear my throat once. That’s all it takes.
Eyes lift. Shoulders square. The final stragglers still their movements. I take a moment to scan the tiered rows—not to count heads, but to observe patterns. Who sits too rigidly, who slouches with false confidence, who’s already pretending not to care.
Attendance matters on the first day. Not because it’s mandatory—Oxford doesn’t coddle—but because absence, like presence, is always a choice. And choices reveal priorities.
“Before we begin,” I say, keeping my tone neutral, “introduction first,” my voice echoes around the now-silent lecture hall as I speak into the lavaliere. “This is Alistair Hertley, your professor for the Michaelmas Term on Advanced Systems & Signal Interpretation.”
A faint ripple of surprise passes through the room. Oxford professors don’t usually bother with introductions. Names are administrative details. But this is the first session of Michaelmas that seems different. And unfamiliar variables require classification.
“I don’t need autobiographies, of course. Your name, your course, and whether this is your first time in the department.” I add because some students tend to repeat the course because of its level of difficulty. “Start from the left.”
A few students shift, glancing at one another before the first obliges. Names follow. Predictable ones—engineering, physics, double honors, some as expected—either repeat or retake.
Mostly final-year undergrads with some in their third year. Some are trying desperately hard to sound confident, others are resigned to exposure. I catalogue accents, posture, and cadence too.
Halfway through the room, I hear a soft, feminine voice.
“Ariella Keinston,” she says with the perfect amount of confidence, as if she’s born to be a Keinston.
The name lands wrong. It’s recognizable in a way most names aren’t. A surname that carries weight beyond the academic register. Corporate boardrooms and discretion clauses. Mainly blood-clean money disguised as innovation.
I lift my gaze fully for the first time since introductions began.
She’s seated two rows up, posture deceptively relaxed, hands folded loosely on the desk in front of her. She not only looks but also sounds younger than the rest—not in a way that reads na?ve, but deliberate and real.
“Electronics and Electrical Engineering. First year. First term.”
That earns her the room.
My eyebrows arch up a little. It’s unlikely to have such an undergrad enrolled in my classes. Keinstons do have it easy.
A murmur ripples through the rows—barely restrained, quickly silenced by my presence. First-year students don’t sit in my lectures. Not without exceptional justification or someone moving pieces behind the scenes.
My expression doesn’t change. Internally, however, something shifts—not surprise, exactly. Interest.
“You’re too early, Miss Keinston,” I say at last, voice smooth, almost amused.
“I prefer not to wait, Mr. Hertley.”
That innocent voice of hers carries a gentle challenge underneath.
What catches my attention is that her accent doesn’t sound Australian at all, even though Keinstons are native there. She sounds more like a British girl.
I nod my head once, acknowledging her response without endorsing it. “Noted.”
Introductions resume, but the rhythm is altered now. They’ve sensed the imbalance amongst them.
By the time the last name is spoken, the room suddenly feels more alert.
I allow myself a smile as I turn back to the display.
It makes people comfortable. That warmth disarms suspicion. Men like me need to appear human, so the world trusts us with its secrets. So I perfected it years ago.
“Very well then,” I say pleasantly and push my glasses up my nose bridge. “Let’s begin.”
I wake the display of my iPad and connect it to the digital board. The title of the module illuminates the screen—ADVANCED SYSTEMS AND SIGNAL INTERPRETATIONS. Michaelmas Term.
“This course is not designed to be forgiving, just as the name implies,” I begin evenly. “If you’re here because you believe intelligence alone is sufficient, you’ll struggle. If you’re here because you think prior success guarantees future competence, you’ll fail,” I pause deliberately.
“And if you’re here because you enjoy being challenged, then you may find this term instructive,” my eyes instinctively swipe over to the youngest student here and find her listening to me with rapt attention—Ariella.
I turn back to the board and begin writing without ceremony, letting the system take shape beneath my hand as though it were nothing more than a routine exercise.
Clean vectors emerge first, followed by familiar operators and transfer functions—structures every senior in this room should’ve seen before in diluted form.
I speak as I write, explaining each component with the ease of repetition, outlining assumptions that sound reasonable enough to pass unchallenged.
“This,” I say, gesturing toward the diagram, “is a closed-loop feedback system under nominal conditions. Input signal, amplification, corrective feedback.”
Then, without emphasis, I add another layer.
A recursive element slips into place, almost polite in its subtlety. The sort of addition that appears harmless until it compounds, until it feeds on itself quietly enough that no one notices the damage being done.
I continue evenly. “It’s a common architecture used everywhere from communication networks to biomedical instrumentation. When tuned correctly, it suppresses instability. When mishandled…” I let the sentence trail off, unfinished by design.
Pens move faster. Screens glow brighter. Heads nod with a confidence that has not yet earned its place.
I proceed as though the system behaves exactly as advertised, defining parameters, assigning values, discussing convergence and signal-to-frequency ratios with the calm assurance of inevitability.
I don’t mention phase drift or correct the assumption about frequency independence. And I certainly don’t acknowledge what happens when an error is allowed to feed back into recursion unchecked.
Because the flaw is already buried there within the structure. It’s a feedback delay mismatched to the recursion depth, which might look trivial on paper but is catastrophic in practice.
Senior students should notice that.
To my disappointment, they don’t.
I can tell by how quickly they write, or from the absence of hesitation, the way they’re accepting the conclusion I’m guiding them toward. They’re following the path I laid out for them, unaware that it leads somewhere unsafe.
I extend the model further, letting the recursive loop run freely, and then ask, almost casually, “What do we expect the output to do after sufficient iterations?”
A few hands begin to rise. I don’t acknowledge them right away because I notice one particular student is looking at me with untamed confidence in her hazel eyes. Ariella.
The moment I look at her, she joins the rest of the students by raising her hand subtly.
I allow my attention to linger on her for a fraction longer than necessary, careful not to betray the observation. She wants only me to notice.
At last, I gesture toward the raised hands. “You,” I say, selecting one at random.
A final-year student straightens immediately, confidence locking into place as he delivers his answer with polished assurance. He speaks, gesturing toward the board as if the system has already resolved itself in his favor.
I let him finish before dismantling the response with careful precision. Then I point out the unverified assumption, alter a single variable, and explain how attenuation becomes amplification once the error term feeds back into the recursion.
His certainty drains visibly as the structure collapses beneath his logic.
Another student follows.
Then another…
Each answer is being delivered confidently. Most incorrect. Some almost correct.
The room grows quieter with every correction, unease replacing certainty as pens slow and eyes flick away from the board in search of reassurance.
“Miss Keinston?” I point at her at last, making her the center of attention once again for the day.
The room stills almost immediately. Several heads turn in her direction, curiosity sharpening into something closer to disbelief.
“Uh—yes,” her eyes flick briefly toward the board before returning to me, already assembling the structure of her response. “The issue isn’t convergence, it’s the assumption that the system remains memoryless once recursion is introduced.”
A ripple moves through the room, not as noise but as surprised attention recalibrating.
She continues without looking at anyone but me.
“You’ve defined the noise as uncorrelated, but the feedback loop reintroduces it with a delay.
That creates dependency across iterations.
The error term isn’t dissipating, it’s accumulating.
” Which means stability only exists in the short term,” she finishes. “After that, the system diverges.”
I don’t interrupt her to correct her phrasing. I don’t step in to reclaim control of the room.
Instead, I slowly turn back to the board.
I redraw a single segment of the model—adjusting the feedback delay, isolating the frequency path, and making visible the propagation she just described.
The structure responds immediately, just as I assumed. The flaw surfaces as cleanly as if it had been waiting for permission to be acknowledged.
She hasn’t just answered the question. She’s reframed the problem.
A few students shift in their seats now, their earlier certainty evaporating as they stare at the board with dawning comprehension.
Others look at her, then back at the system, recalculating their understanding of both.
I set the stylus down.
“That’s the correct interpretation, thank you, Miss Keinston.”
She smiles shyly, not triumphantly.
I turn back to the class, already moving on, already folding the corrected model into the broader framework of the lecture.
But internally, I know Keinston’s money didn’t buy that intelligence. Someone taught her. Or more troubling still, she taught herself.
Either way, she isn’t an accident. She’s a dangerously unclassified variable in my equation.