How to test the accounts we build about ourselves

Years after an argument, you may still be able to describe the room, the remark that changed the atmosphere and the moment you decided the other person had been unfair.  You might even have the messages that followed.  Yet the part of the story most resistant to examination is often the sentence that joins those details together: “I was trying to resolve it, and they wanted a fight.”

Perhaps that is true.  Perhaps they were trying to resolve something too, but you disagreed about what needed resolving.  Perhaps the remark was unkind and your account of what led to it is incomplete.  The difficulty is that you do not experience your explanation as one interpretation among several.  You experience it as what happened, particularly after you have told it often enough to make it part of the history of the relationship.

This series began with the limits of perception, followed an event into memory, and watched memory become a story about the self and other people.  It then moved into the accounts groups carry forward as history.  The final question is practical and uncomfortable.  When the narrator is also a participant, what would it take to test the story without pretending we can step outside ourselves altogether?

The answer begins with a modest distinction.  We can be certain that an experience mattered to us without being certain about every explanation we later attached to it.  We can remember an injury accurately and still misread a motive.  We can discover that another person experienced the same event differently without deciding that either account must be invented.  What we need is a way of examining a claim that leaves room for both evidence and human perspective.

The account does work for us

A story about the past is rarely just a conclusion about the past.  It may tell us why a marriage failed, why we left a profession, why a parent behaved as they did, or why an organisation never listened.  Change one explanation and other conclusions may need to move with it.  This is why an apparently small factual challenge can feel disproportionate.  The challenged detail has become a support for a much larger structure.

Psychologist Ziva Kunda's influential account of motivated reasoning proposed that our goals can affect which beliefs and strategies we use to evaluate a question.  Wanting accuracy and wanting a particular conclusion are different motives, and both can operate in the same person.  The argument does not imply that we can simply believe anything convenient.  A preferred explanation usually still has to be plausible enough to defend, which is one reason intelligence can sometimes supply a more elaborate defence rather than a better answer.

There is a complication worth keeping.  A recent review of motivated reasoning notes continuing uncertainty about when and why it occurs.  It is easier to bend ambiguous information than an unambiguous record, and strong incentives to get something right can alter how evidence is considered.  Nor is every disagreement with us proof that someone else is protecting their identity.  That accusation has a convenient way of exempting its author from the very scrutiny they recommend.

So the starting point is not to suspect every attachment we have to our own account.  It is to ask what the account allows us to preserve.  Does it make a difficult decision seem necessary?  Does it keep a loved person wholly good or wholly bad?  Does it explain a failure without requiring us to revisit a choice?  Those questions do not prove the story false.  They reveal why it might be hard to judge.

Find the point where the facts become a story

Return to the argument.  The other person said, “You never listen.”  You replied, “That is unfair.”  You left the room before the discussion ended.  If those details can be established, they form a useful starting record.  “They wanted a fight” belongs to a different category.  It attributes an intention.  “I was trying to resolve it” also attributes one, this time to ourselves, and may leave out the effect of our behaviour on the other person.

This distinction can sound pedantic until two people begin disputing an event.  They may agree on the words and disagree about whether the words were sarcastic, desperate, dismissive or merely clumsy.  A pause might have felt like contempt to one person and restraint to the other.  Even a contemporaneous message can establish that something was written, without establishing the writer's full intention or the effect it had when received.

The useful question is where we crossed from observing a behaviour to claiming to know its cause.  If we say, “She excluded me because she was threatened by me,” what is the observed exclusion, and what evidence bears on the reason?  If we say, “I had no choice,” which alternatives were available at the time, including ones we rejected because of their cost?  An explanation may survive those questions.  It will be stronger for having answered them.

Chronology matters as well.  We often remember an episode through what it eventually became.  A friendship that ended painfully can acquire warning signs in every early disagreement.  A venture that succeeded can make earlier doubts look like obstacles that wise people overcame.  To recover some of the uncertainty that existed then, read the messages, notes or minutes from before the outcome was known.  A diary entry is not an impartial transcript, and an email can be strategic.  Its advantage is narrower: it shows what someone recorded at a particular time, before a later ending reorganised the story.

An incomplete record should remain incomplete.  Silence in an archive does not prove consent; an omitted reply does not prove that no reply was given.  The same discipline we demand of memory should apply to documents.  They are evidence to interpret, with their own authors, purposes and gaps.  Their value lies partly in their ability to interrupt a seamless retrospective account.

Invite a version you cannot control

The next source of friction is another person's account.  We should choose that person with care.  Someone who witnessed the event may know details we missed, but have interests of their own.  Someone close to us may know our usual patterns, but may also feel pressure to protect us.  And the person with whom we disagree may remember their motives more clearly than we can, while knowing no more than we do about ours.

The point is not to convene a tribunal and count the witnesses.  It is to find out which parts of our story survive contact with information we did not select.  A useful invitation is specific: “This is what I remember saying.  What do you remember hearing?”  Or: “I have always understood your decision as a response to that meeting.  Is that how it looked to you then?”  The answers may still be self-serving, but they can reveal a missing sequence, a different constraint, or a meaning that never occurred to us.

Listening to another account is especially difficult when it contains a judgement of our character.  A person can be mistaken about our intentions and accurate about the harm they experienced.  We can be accurate about our intentions and mistaken about the impression our actions created.  Those are compatible propositions.  Many disputes become intractable because each participant treats acceptance of one proposition as surrender of the other.

There is a limit to this method.  The person who harmed us is not automatically entitled to set the terms of an inquiry into what happened.  Access to their version may be impossible, unsafe or unhelpful.  Sometimes the evidence supports a firm conclusion already.  Intellectual humility does not require endless reopening of a well-supported judgement, particularly when doing so repeatedly serves someone else's interest in avoiding responsibility.

Give the other explanation a fair test

Suppose we believe a colleague blocked our proposal because they disliked us.  What else could explain the same behaviour?  They may have thought the proposal was weak, faced a constraint they had not disclosed, protected their own project, or simply handled disagreement badly.  Generating alternatives is easy.  The harder part is asking which evidence would distinguish them.

An older experimental line of work on considering the opposite found that deliberately entertaining a contrary possibility can reduce some judgement biases in controlled tasks.  That finding does not make the exercise a reliable cure for every personal dispute.  It does suggest why merely repeating “I might be biased” is inadequate.  We need to develop an alternative account enough to see what it would predict.

If the colleague's objection really concerned a deadline, we might expect the same objection to appear in other proposals.  If the issue was personal, we might look for a pattern across comparable decisions, while resisting the temptation to classify every ambiguous interaction as confirmation.  If evidence supports several explanations, the most honest conclusion may be more limited than the one we started with: “They opposed my proposal, and I do not know exactly why.”  That sentence can be irritatingly unsatisfying.  It may also be true.

The strongest test is to decide in advance what would change our mind.  Would a contemporaneous note matter?  Would a consistent account from an independent witness?  Would discovering that we had previously described the event differently?  If we cannot name anything, the explanation has stopped functioning as a claim we can examine.  It may still express something real about how we feel, but it is now doing different work.

This principle applies to institutions as much as individuals.  The Columbia investigation in the previous article showed how a history of safe returns helped shape the interpretation of a new hazard.  An organisation can record dissent and still have no effective route for it to alter a decision.  The practical test is not whether a meeting allowed someone to disagree.  It is whether the reasons for disagreement were investigated at the point when the decision could still change.

What an honest revision looks like

Sometimes the evidence alters one part of a story and leaves its central meaning intact.  You may discover that a remark was made later in the evening than you recalled, without changing your judgement that it was cruel.  Sometimes the event remains certain but its meaning changes.  A parent's absence may have been real even if a newly discovered record complicates what they intended.  A good revision makes that distinction explicit, rather than using uncertainty about one detail to erase everything or treating every challenge as an attack on the whole.

Revision also need not produce an attractive new identity.  The popular language of “rewriting your story” often suggests that the goal is to find a more empowering version of events.  That might be valuable for some purposes, but it is a different task from asking what the evidence supports.  A person can replace a story in which they were always the victim with one in which they were always responsible, and still be arranging the past around a totalising role.  The most accurate account may contain injury and agency, good motives and harmful effects, insight in one episode and stubbornness in the next.

There are costs to demanding perfect clarity.  Most ordinary lives leave patchy records, and people who witnessed an event may be gone, estranged or unable to give a useful account.  The past will not yield a definitive verdict on every relationship.  We can nevertheless be precise about the boundary of our knowledge: what happened as far as we can establish, what we think it meant, and what we cannot know.  That boundary is not a failure of the inquiry.  It is one of its findings.

The series began with the question of how accurately we see what is before us.  Its answer has never been that perspective makes truth impossible.  Perception is selective; memory reconstructs; identity and group life influence which account survives.  Yet these processes also leave traces we can examine, claims we can compare and decisions we can revisit.  We get closer by allowing an unwelcome fact to matter, even when it does not tell us exactly what to believe instead.

The hardest part is not admitting, in the abstract, that our story could be wrong.  It is recognising the moment when a specific piece of evidence asks us to change a sentence we have used to explain our life.  We may keep the sentence after all.  But first, we have to let the evidence speak.