Essays asking students to analyze the Underground Man's psychology tend to fall into a common trap: cataloguing his contradictions, his simultaneous claims to superiority and self-loathing, his desire for connection paired with active sabotage of it, without ever moving toward a unified explanation. This is understandable given how genuinely contradictory the character is, but a strong essay needs to argue for some coherent psychological logic underneath the surface chaos. Grading these essays well means recognizing the difference between thorough cataloguing and genuine psychological analysis, since students often conflate the two. Feedback that names this distinction clearly tends to produce the most improvement in subsequent drafts.

A useful grading question is whether the essay offers a "because" statement that connects multiple contradictions to a single underlying cause. An essay that argues the narrator sabotages his relationship with Liza because genuine intimacy would threaten the intellectual superiority he depends on for self-worth is offering exactly this kind of unifying explanation. An essay that simply lists the sabotage as one more example of his general inconsistency has stopped short of the real analytical work. Marking this distinction directly, rather than assuming students will infer it, helps clarify what "deeper analysis" concretely requires on this particular assignment.

It also helps to watch for essays that diagnose the narrator using modern psychological terminology, like "narcissism" or "depression," without grounding that diagnosis in the text's own language and context. This is not inherently wrong, since modern readers naturally bring contemporary frameworks to older texts, but the strongest essays explain how the text itself supports the diagnosis rather than simply applying a label. A student who argues the narrator displays patterns consistent with what we now call self-sabotaging behavior, and supports that claim with specific textual moments, is doing careful work. A student who applies the label without that textual grounding has skipped a necessary step in the argument.

Moving Students From Cataloguing to Argument

The most effective feedback for essays stuck in cataloguing mode is a direct question that forces synthesis: "you have shown me five contradictions; if you had to explain all five with one sentence about what the narrator actually wants, what would that sentence be?" This question respects the genuine effort students put into finding contradictions while pushing them toward the harder work of synthesis. It also gives graders a fast, repeatable comment that can be adapted slightly to each essay's specific examples without requiring an entirely new comment from scratch each time. This kind of targeted question tends to unlock real improvement more reliably than a general note to "analyze further."

• Check for a clear unifying explanation connecting multiple contradictions, not just a list
• Require modern psychological terms to be grounded in specific textual evidence
• Ask students to synthesize their observations into one governing claim about motivation
• Reward essays that trace how the narrator's psychology shifts across the two parts
• Flag diagnosis-style language used without textual support as an area for revision

Handling Essays That Read the Narrator Too Sympathetically or Too Harshly

Student essays on this novel often skew toward one of two extremes: excessive sympathy that excuses the narrator's cruelty as simply the product of trauma, or excessive judgment that treats him as irredeemable without examining why Dostoevsky built in moments of genuine vulnerability. Both readings can be defended with textual evidence, and neither should be penalized outright, but essays that acknowledge the tension between sympathy and judgment tend to be more sophisticated than essays that commit fully to one extreme. Feedback here should push students to grapple with the discomfort the novel is designed to create, rather than resolving that discomfort too quickly in either direction. This is one of the harder balances to teach, since the instinct to pick a clear side is strong.

One helpful technique is asking students directly whether the narrator himself seems aware of the contradiction they are analyzing, since his self-awareness, or lack of it, is often central to a sophisticated psychological reading. A narrator who knows he is being cruel and does it anyway is psychologically different from one who genuinely believes his own rationalizations, and essays that address this distinction tend to show real depth. This question also naturally pushes students away from simple moral judgment and toward the more interesting question of what the narrator's self-awareness, or blindness, reveals about human psychology more broadly.

Grading Psychological Analysis Essays Efficiently

Because these essays require tracking psychological consistency across an entire novel, they can be more time-consuming to grade than essays focused on a single scene or symbol. A rubric that explicitly separates "identification of contradiction" from "synthesis into unified explanation" helps graders move through essays more quickly by locating exactly where an essay stops short. AI-assisted feedback tools trained on this same distinction can flag essays that remain in cataloguing mode, giving instructors a fast way to prioritize which essays need the most substantial written feedback. This targeted approach saves meaningful time without lowering the bar for the psychological depth expected on this assignment.

Keeping a short bank of strong synthesis statements from past semesters, shared with students as models rather than as answers to copy, also helps set expectations before the essays are written. Seeing what a real "because" statement looks like, connecting multiple contradictions to one underlying psychological logic, gives students a clearer target than an abstract description of the skill ever could. Over time, this kind of modeling tends to reduce the number of essays that stall at the cataloguing stage, which in turn reduces the amount of redirection feedback needed during grading itself.