Ian McEwan's novel "What We Can Know" asks whether a person can ever truly know another mind, and the answer the book dramatizes is that knowledge of other people is always inferential, provisional, and mediated by memory and narrative. McEwan builds that argument through a near-future setting, a long marriage, and a literary mystery, using the mechanics of fiction to test the limits of first-person certainty.
The Mechanism of Limited Knowledge
McEwan's central mechanism is a narrator who reasons carefully but cannot escape his own vantage point. The novel presents knowing as a process of building models from incomplete evidence, not as direct access to facts. That distinction maps onto a real philosophical problem: other minds are never observed, only inferred from behavior, language, and consistency over time.
The book uses a long marriage as its laboratory. Two people who have shared decades still hold private interiors the other cannot verify. McEwan dramatizes this by letting the reader see the narrator's confidence and his blind spots at once. The technique produces a specific effect: the reader knows more than the narrator about his own limits, while still knowing less than the novel's full truth.
Memory compounds the problem. Recollection is reconstructive rather than archival, so a person's account of the past is a rebuilt version, not a recording. McEwan treats memory as evidence that has been handled, and handling leaves marks. The novel's mystery therefore turns less on hidden facts than on which accounts can be trusted and why.
Plainly: the book argues that certainty about another person is a feeling, not a proof.
Quantified Context: How Much Can Anyone Know
The novel's implicit claim is that even intimate knowledge is partial. Research on relationship knowledge supports the general shape of that claim without supplying exact figures the book cites. Partners are more accurate than strangers at predicting each other's preferences, but accuracy drops sharply for novel situations and for emotional states under stress.
Long relationships improve prediction of habits and routines more than prediction of private reactions. That asymmetry is exactly what McEwan stages: the marriage knows its rituals precisely and its hidden interiors poorly. The narrator's expertise is real but bounded, and the boundary is where the plot lives.
McEwan's near-future frame adds a second layer. Technologies that promise more data about a person do not automatically produce more understanding, because interpretation still requires a model. More information can raise confidence faster than it raises accuracy, which is a documented pattern in judgment research. The novel treats that gap as its engine.
The numbers that matter here are relational, not statistical: decades of shared history on one side, unverifiable interior experience on the other, and a narrator who mistakes fluency for truth.
Edge Cases: Where Knowledge Breaks Down
McEwan tests knowing at its weakest points. One is the case of a person who is highly articulate about themselves; fluency reads as self-knowledge but can be performance. Another is the case of a person who is silent or reserved; the absence of narration reads as emptiness when it may be depth.
A third edge case is retrospective certainty. Someone who reconstructs a past event confidently may be reconstructing a story rather than retrieving a fact. The novel uses this to destabilize the narrator without declaring him dishonest. He is not lying; he is remembering in a way that protects a version of himself.
A fourth is the observer's stake. When a person has something to lose from a conclusion, the conclusion tends to arrive early and feel obvious. McEwan builds this bias into the narrator's reasoning so the reader can watch certainty form before the evidence supports it.
The counterintuitive detail: the novel suggests that greater intimacy can reduce accuracy, because close relationships generate strong priors that override new evidence. Strangers must look; intimates assume.
Misconceptions About the Novel and Its Argument
A common misconception is that the book is primarily about artificial intelligence or surveillance technology. Those elements frame the problem, but the core subject is ordinary human knowledge between people who love each other. The technology is a pressure device, not the thesis.
Another misconception is that the novel endorses radical skepticism, the view that nothing can be known. It does not. The narrator reaches real conclusions, some of them correct. The book's position is fallibilism: knowledge is possible but always revisable, and confidence should track evidence.
A third is that the mystery is solved by a single revelation. McEwan's plots typically resist that structure. Understanding arrives as adjustment and accumulation, and the reader's revised model of a character matters more than a climactic fact.
A fourth is that the narrator is unreliable in the simple sense of deceiving the reader. He is unreliable in the way all first-person narrators are: he reports what he notices, and he notices what his model of himself permits.
Treat narrator confidence as a claim, not a finding. Distinguish what a character reports from what the novel shows. Watch for scenes where the narrator's model of a person predicts and then distorts. Note where memory is described as selection rather than storage. Separate the novel's technological frame from its human argument.
Comparing Ways of Knowing
McEwan's novel sits inside a long tradition of fiction about epistemic limits. The comparison below maps the novel's approach against adjacent frames without ranking them.
| Frame | What counts as knowing | Main limit | How McEwan's novel differs |
|---|---|---|---|
| First-person testimony | Coherent account of experience | Confirmation bias, selective memory | Narrator's coherence is treated as suspect |
| Behavioral inference | Consistent patterns over time | Novel situations, masking | Marriage shows pattern knowledge without interior access |
| Documentary record | Verifiable artifacts | Gaps, framing, forgery | Evidence is present but interpretation dominates |
| Technological surveillance | Captured data | Data without model, privacy erosion | More data raises confidence, not understanding |
| Literary omniscience | Full interior access | Not available in life | Used as contrast to the narrator's partial view |
| Scientific psychology | Replicable measurement | Group averages, not individuals | Explains the asymmetry the novel dramatizes |
The table's point is that each frame trades one kind of access for another kind of blindness. McEwan's narrator uses several at once and still misses what the reader can see.
A Timeline of Knowing in the Novel
The novel's epistemic arc follows a recognizable course. The timeline below maps the stages of the narrator's understanding without disclosing plot events.
| Stage | What the narrator holds | What remains hidden | Reader's position |
|---|---|---|---|
| Initial confidence | A settled account of self and marriage | Private interiors, unexamined history | Sees stability as fragile |
| First discrepancy | A detail that does not fit | Whether the detail is significant | Notices the misfit early |
| Model defense | Explanations that preserve the account | Alternative interpretations | Watches rationalization operate |
| Accumulating friction | Multiple small inconsistencies | The pattern connecting them | Assembles the pattern first |
| Revision | A modified model of a person | Whether revision is sufficient | Sees both gain and loss |
| Provisional settlement | A workable, revisable understanding | Residual uncertainty | Recognizes the limit as permanent |
The shape is not a puzzle solved but a model updated. McEwan gives the reader the update before the narrator, which is how the novel teaches its epistemology.
Practical Actions for Reading and for Knowing
The novel rewards specific reading habits. Track where the narrator's confidence rises without new evidence. Mark every scene where memory is described as effortful or selective. Compare what the narrator claims to know with what the novel has shown the reader.
For the real-world question the book raises, a few concrete practices follow from its argument. Record important conversations while they are fresh, because recall drifts within hours and confidence does not. State your confidence separately from your conclusion, so revision costs less. Ask what evidence would change your mind before you finish forming the view.
Write down the specific evidence behind any strong claim about a person. Separate observation from interpretation in your notes. Set a threshold: if new evidence contradicts your model twice, revise the model. Check whether intimacy is producing assumption rather than attention. Treat fluency and certainty as signals to look harder, not to stop.
Thresholds matter because they prevent drift. If your account of a person survives two independent contradictions unchanged, the account is doing identity work, not truth work. That is the novel's warning, and it applies outside fiction.
Limitations of the Novel's Argument and This Analysis
McEwan's case is built from a single narrator in a single marriage, which is a powerful but narrow sample. The novel cannot establish how often intimacy reduces accuracy, only that it can. Treating its claims as general findings would exceed the evidence the fiction provides.
The same caution applies here. Claims about memory reconstruction and about confidence outrunning accuracy are supported by research traditions, but this article does not cite specific studies and does not invent figures. Where numbers would be required for proof, the argument stays qualitative.
The novel also cannot settle the philosophical question it dramatizes. Whether another mind is knowable in principle is unresolved in philosophy, and McEwan wisely does not resolve it. He shows what provisional knowing looks like in practice, which is more useful than a verdict.
There is one genuine tension in the book worth naming. If all knowledge of others is model-building, then the narrator's final understanding is itself a model, subject to the same critique he applies to everyone else. McEwan does not exempt him. That consistency is the novel's strongest philosophical move.
Readers who want a clean answer will find the book frustrating. Readers who want a demonstration of how knowledge actually works under uncertainty will find it precise. The distinction between those two readers is itself a small experiment in what we can know.