Short Answer

Learn how reliable eyewitness accounts are, why memory can change, and how historians test firsthand testimony using context and corroboration.

A soldier remembers an enemy attack beginning before sunrise.

Another soldier from the same unit insists that daylight had already broken.

A civilian remembers thousands of people filling a city square.

A police report estimates only several hundred.

A political leader later recalls having opposed a controversial decision from the beginning.

Private correspondence written at the time suggests otherwise.

Which witness should we believe?

Eyewitness accounts are among the most compelling sources historians encounter because they provide testimony from people who were actually present. But firsthand observation does not create a perfect recording of the past.

Human beings do not experience events like cameras.

We notice some things and overlook others. We interpret what we see. Memories fade. Later information can influence recollection. People sometimes confuse what they witnessed with what they subsequently heard. And when memories are recalled years later, present-day knowledge can shape how earlier experiences are understood.

Scientific research on eyewitness memory has documented inherent limitations in perception and memory, including the possibility that recollections can be reconstructed, updated, and distorted after an event.

Yet this does not mean eyewitness accounts are useless.

Far from it.

A firsthand account may preserve information available nowhere else. It can reveal what people experienced, noticed, believed, feared, expected, misunderstood, and remembered.

The historian’s task is therefore not simply to decide whether an eyewitness is reliable or unreliable.

A better question is:

Reliable for what?

How Reliable Are Eyewitness Accounts? The Short Answer

Eyewitness accounts can be highly valuable historical evidence, but their reliability varies.

Their accuracy can depend on:

how close the witness was to the event,
how long they observed it,
visibility and environmental conditions,
stress and distraction,
what the witness already knew,
how quickly the experience was recorded,
exposure to later information,
conversations with other witnesses,
the questions subsequently asked,
the witness’s purpose and intended audience,
the amount of time between the event and recollection.

Research on eyewitness identification has found that conditions such as poor lighting, limited viewing time, stress, and distraction can affect perception, while subsequent events can affect memory.

For historians, the practical conclusion is straightforward:

Treat eyewitness testimony as evidence to investigate, not as a transcript of reality.

Eyewitness Does Not Mean Omniscient

The word eyewitness can give an account unusual authority.

After all, the person was there.

But being present is not the same thing as seeing everything.

Imagine a demonstration involving 20,000 people spread across several streets.

One witness stands near the main stage.

Another stands beside a police barricade three blocks away.

A third watches from an apartment window.

A fourth arrives after clashes have already begun.

All four are eyewitnesses.

Their accounts could differ dramatically without anyone deliberately lying.

Every witness occupies a particular position in space and time.

Historians therefore begin with basic questions:

Where was the witness?

What could they actually see?

How long were they present?

What happened outside their field of vision?

These questions often matter more than whether the witness sounds confident.

Eyewitness Memory Is Not a Video Recording

One of the most important findings from modern memory research is that recollection is not simply a stored movie that a person plays back later.

For a witness to report an event accurately, information must first be sensed and perceived, then encoded and stored, and later retrieved. Each stage has limitations. The National Academies’ extensive review of eyewitness science emphasizes that accuracy depends on the limits of both perception and memory.

This distinction matters enormously for historical sources.

Suppose someone writes fifty years after witnessing an assassination:

I can still see the scene perfectly.

The witness may sincerely experience the memory as vivid.

But vividness alone does not establish that every detail has remained unchanged for fifty years.

The witness is remembering.

And remembering is an active cognitive process.

A Witness Can Be Completely Sincere and Still Be Wrong

Historians should distinguish between two very different questions:

Is the witness lying?

and

Is the witness accurate?

A person can honestly describe something that did not happen exactly as remembered.

This is especially important when working with:

diaries,
memoirs,
oral histories,
interviews,
courtroom testimony,
veterans’ recollections,
survivor testimony.

Errors do not necessarily indicate deception.

A witness may confidently remember a conversation taking place on Tuesday when records show it happened on Thursday.

They may merge two similar events.

They may attribute something to the wrong person.

They may later incorporate information learned from newspapers or other witnesses.

Historical source criticism must therefore go beyond judgments about personal honesty.

Sincerity and accuracy are not the same thing.

Perception Comes Before Memory

Memory cannot preserve information that was never perceived accurately in the first place.

Consider a battlefield.

There may be:

smoke,
darkness,
explosions,
rapid movement,
extreme noise,
confusion,
physical danger.

A soldier might genuinely believe that thousands of enemy troops were advancing.

From a limited position, however, the soldier may have observed only one section of a much larger operation.

Likewise, a civilian witnessing a nighttime disturbance might misidentify uniforms, weapons, vehicles, or individuals because visibility was poor.

The National Academies identifies conditions such as dim lighting, brief exposure, stress, and visually distracting objects as factors that can affect eyewitness perception.

For historians, this leads to a crucial question:

Was the witness actually in a position to observe what they later claimed?

Stress Can Affect Eyewitness Memory

Historical events worth recording are often stressful.

Battles.

Riots.

Natural disasters.

Executions.

Assassinations.

Bombings.

Mass displacement.

Political violence.

A witness experiencing such an event may have been frightened, injured, exhausted, hungry, or overwhelmed.

A major meta-analysis examining dozens of experimental tests found substantial evidence that high levels of stress can negatively affect both eyewitness identification and recall accuracy.

This does not mean:

stress = false memory.

Nor does it mean every traumatic memory is unreliable.

Stress interacts with attention, circumstances, the type of information being remembered, and other variables.

The more defensible historical principle is:

Extreme circumstances do not automatically make a witness more accurate simply because the experience was unforgettable.

An event can be unforgettable while some details remain uncertain.

Central Details and Peripheral Details May Be Remembered Differently

Imagine witnessing a dramatic explosion.

You may remember very clearly:

the explosion happened.

But perhaps not:

what color a nearby vehicle was,

exactly where each person stood,

the precise sequence of minor actions,

or the clothing worn by someone at the edge of the scene.

Research on eyewitness memory suggests that memory for different elements of an event can vary considerably. In one experimental study involving a distressing event, participants performed differently on central and peripheral details, while misleading information affected later recollection.

Historians should therefore avoid treating a witness’s account as having one universal reliability score.

A witness may be very reliable about the central event:

There was an explosion.

and much less reliable about:

The man beside the doorway wore a dark blue coat.

Reliability can vary within the same testimony.

Memory Changes With Time

Time creates another problem.

A diary written the evening after an event and a memoir written forty years later are both potentially valuable primary sources.

But they preserve different forms of testimony.

The diary records the writer’s understanding close to the event.

The memoir records what the event became in the writer’s memory.

Experimental research has repeatedly found that delay can affect recall. For example, research examining eyewitness memory across immediate and delayed tests found reduced accuracy after longer retention intervals, while other work has shown that delay can increase vulnerability to misleading post-event information.

For historians, the date of creation matters almost as much as the date of the event.

Always ask:

When was this account actually recorded?

Contemporary Accounts Are Not Automatically Accurate

It would be tempting to solve the problem by trusting only accounts written immediately after an event.

But contemporary accounts have their own limitations.

Someone writing immediately may:

lack broader information,
believe rumors,
misunderstand what just happened,
react emotionally,
know only what occurred nearby.

Suppose a soldier writes after a battle:

We destroyed the entire enemy army today.

A later memoir might correct this exaggeration after the writer has learned what actually happened.

The immediate account may therefore be worse for reconstructing the overall battle.

But it may be much better evidence for understanding:

what the soldier believed that evening.

This demonstrates one of the central principles of historical evidence:

A source can be unreliable for one question and extremely reliable for another.

Later Recollections Have Hindsight

A memoir introduces a different problem:

the witness knows how the story ended.

Imagine a politician remembering the months immediately before a revolution.

At the time, nobody knew whether the government would survive.

Thirty years later, the politician knows exactly what happened.

That knowledge can subtly reorganize memory.

Early warnings may appear more obvious.

Certain people may seem more important.

Events that eventually proved decisive may receive greater emphasis.

Alternative possibilities that seemed realistic at the time may disappear from the narrative.

Historians call attention to this because the past was not experienced with knowledge of its future.

The American Historical Association emphasizes the importance of historical context and contingency: people living through events did not know the outcomes historians now take for granted.

The Misinformation Effect

One of the most important discoveries in eyewitness research concerns what happens after an event.

People encounter new information.

They read news reports.

They speak with other witnesses.

They see photographs.

They hear rumors.

They watch documentaries.

They read books.

Sometimes this later information becomes mixed with the original memory.

Research on what psychologists call the misinformation effect has shown that exposure to misleading post-event information can impair later reports about what someone originally witnessed.

This phenomenon has obvious implications for historians.

Imagine interviewing a veteran fifty years after a famous battle.

During those fifty years, the veteran may have encountered:

regimental histories,
films,
commemorations,
other veterans’ stories,
newspaper retrospectives,
historical books.

The resulting testimony may contain a mixture of:

personal memory

and

later acquired knowledge.

Separating those components can be difficult.

Witnesses Can Influence One Another

Suppose five people witness a political assassination.

Immediately afterward, they gather together.

One says:

I definitely heard three shots.

Another initially remembers only two.

After repeated discussion, that second person begins remembering three as well.

Experimental research has documented memory conformity, in which information obtained from another witness can influence someone’s later account.

This has an important consequence for historical corroboration.

If five witnesses eventually tell the same story, historians should ask:

Did these accounts develop independently?

Five independent witnesses agreeing is powerful evidence.

Five people repeating a story after discussing it together is less independent than it appears.

Repetition Does Not Necessarily Mean Independent Confirmation

This principle extends beyond witnesses.

Imagine that ten newspapers from 1914 report the same dramatic statement.

That appears to be ten sources.

But investigation reveals:

eight copied a wire-service report,
one copied another newspaper,
only one had an original correspondent.

You do not really have ten independent accounts.

You have one information chain repeated ten times.

The same problem can occur with eyewitness testimony.

Memoirs may repeat famous stories.

Veterans may influence each other’s recollections.

Communities can develop shared narratives.

Historical reliability therefore depends not only on how many sources agree, but on whether those sources are genuinely independent.

Suggestive Questions Can Affect Testimony

How testimony is collected also matters.

Compare these questions:

What did you see?

and

How fast was the red car traveling when it struck the other vehicle?

The second question already supplies information:

there was a car,

it was red,

and it struck another vehicle.

Research on eyewitness memory has demonstrated that post-event suggestion can influence later recollection, which is one reason modern investigators attempt to reduce suggestive interviewing practices.

Historians using oral histories should therefore examine not only the answers.

They should examine the questions.

Who conducted the interview?

What did the interviewer already believe?

Were questions leading?

Were topics suggested?

Was the interview edited?

An oral history is partly a product of an interaction between interviewer and narrator.

What About Witness Confidence?

A witness says:

I am absolutely certain. I would recognize him anywhere.

Should confidence make historians trust the testimony?

The answer is more complicated than the familiar claim that confidence and accuracy are completely unrelated.

Research reviewed in Psychological Science in the Public Interest found that, under carefully controlled initial identification conditions—for example, an uncontaminated first test using fair procedures—high initial confidence can provide meaningful information about accuracy. But that relationship can deteriorate when procedures are biased or memory has already been contaminated.

This nuance is important.

A historian should not reason:

The witness sounds confident, therefore the witness must be right.

But neither should the historian assume:

Confidence never tells us anything.

Instead ask:

When was that confidence recorded?

Confidence expressed immediately after an event or an uncontaminated first recollection is different from confidence displayed decades later after repeated retelling.

Confidence Can Grow Without the Memory Becoming More Accurate

People sometimes become more certain after receiving confirmation.

Research on eyewitness identification has shown that confirming feedback can inflate witness confidence, including confidence in incorrect identifications.

Historical testimony can experience an analogous process.

Imagine someone tells a childhood story:

I remember seeing the president drive past our house.

A relative responds:

Yes! I’ve always heard that you did.

The story is repeated at family gatherings.

Eventually it becomes a defining family memory.

Each repetition may strengthen confidence in the narrative.

That growing confidence does not necessarily mean the historical details have become more accurate.

Famous Stories Can Become Better Than Real Memories

Memory also interacts with storytelling.

Historical witnesses may repeat an experience many times.

Over decades, messy events can become cleaner narratives:

there is a beginning,

a dramatic turning point,

and a memorable conclusion.

Chronology becomes compressed.

Dialogue becomes polished.

Coincidences become meaningful.

This does not necessarily mean the narrator is deliberately inventing the story.

Humans naturally organize experiences into narratives.

For the historian, this creates an important distinction:

A compelling story is not automatically a precise historical reconstruction.

The better the story sounds, the more important it can be to check contemporary evidence.

Bias Is Different From Memory Error

Another distinction matters.

A witness can be inaccurate because memory failed.

But a witness can also shape testimony deliberately or semi-deliberately because of personal interests.

Consider a military commander explaining a disastrous defeat.

The commander may genuinely misremember some details.

But the account may also emphasize information that protects the commander’s reputation.

Similarly, political memoirs may:

minimize mistakes,
exaggerate foresight,
transfer responsibility,
emphasize successes,
omit embarrassing events.

Historians therefore investigate both:

memory

and

motive.

One belongs primarily to cognitive psychology.

The other belongs to source criticism.

Sometimes both operate simultaneously.

Audience Matters

People describe experiences differently to different audiences.

Imagine a soldier writing three accounts.

Private diary

We are exhausted and terrified.

Letter to parents

Everything is fine. Please don’t worry.

Official report

Morale remains satisfactory despite difficult conditions.

Which is reliable?

Possibly all three—in different ways.

The letter is reliable evidence that the soldier wanted to reassure the family.

The diary may be stronger evidence for private emotion.

The report may tell historians how the military officially represented morale.

This illustrates why reliability cannot be separated from purpose.

Ask:

Why was this account created?

Diaries Can Be Wrong Too

Private diaries often feel especially trustworthy because they were not written for future publication.

They can indeed provide extraordinarily valuable evidence.

But private does not mean infallible.

A diarist may:

misunderstand political events,
repeat rumors,
exaggerate crowd sizes,
misidentify people,
remember conversations incorrectly.

A diary’s great strength is often not that every statement is factual.

It is that it preserves what someone thought, experienced, and believed close to the time.

That can be invaluable historical evidence even when particular claims prove incorrect.

Memoirs Need Different Questions

Memoirs are often written years or decades after events.

Instead of asking only:

Did everything happen exactly this way?

historians can ask two questions.

Question 1: What happened?

Here the memoir should be compared with contemporary documents and other evidence.

Question 2: How did the author remember the experience later?

Here the memoir itself becomes evidence of historical memory.

A discrepancy between a wartime diary and a memoir written forty years later can be extraordinarily revealing.

Perhaps the person’s interpretation changed.

That change itself has a history.

Multiple Independent Eyewitnesses Can Be Powerful Evidence

The limitations of eyewitness memory should not lead to excessive skepticism.

Sometimes eyewitness accounts corroborate each other remarkably well.

Suppose researchers find:

a civilian diary,
a soldier’s letter,
an independent journalist’s report,
a government telegram,

all created within twenty-four hours of the same event.

All independently mention that a bridge collapsed at approximately 4 p.m.

The convergence greatly strengthens the conclusion.

Historical thinking requires both skepticism and proportional confidence.

The American Historical Association describes careful historical inquiry as a process of seeking corroboration, assessing competing views, recognizing incomplete records, and remaining open to revising conclusions.

The goal is not to distrust everything.

The goal is to determine how much confidence the available evidence deserves.

When Eyewitness Accounts Conflict

Contradiction is normal.

Suppose three eyewitnesses describe the beginning of a riot.

Witness A

Police attacked first.

Witness B

Demonstrators began throwing objects first.

Witness C

I heard shouting and then saw people running, but I could not see what started it.

It would be tempting to choose A or B.

A historian should first investigate:

where each person stood,
when they arrived,
whether visibility was obstructed,
whether they were participants,
whether they spoke to others before recording their accounts,
when each account was written.

Perhaps Witness A and Witness B stood on opposite sides of a large square.

Both may accurately describe different sequences occurring in different places.

Contradictory testimony does not automatically mean one witness lied.

Sometimes disagreement reveals the complexity of the event.

Historical Accounts Are Partial by Nature

Historians face a broader problem that goes beyond eyewitness testimony.

The historical record itself is incomplete.

The American Historical Association notes that historians construct explanations from partial primary sources that can sometimes support multiple interpretations.

Documents were lost.

Witnesses died without recording their memories.

Governments destroyed records.

Some communities produced far fewer written sources.

Powerful institutions preserved their perspectives much more effectively than marginalized people.

An imperfect eyewitness account may therefore remain extremely valuable if it preserves a perspective otherwise missing from the historical record.

Reliability matters.

So does rarity.

Psychological Research and Historical Sources Are Not Exactly the Same Problem

An important caution is necessary.

Much scientific research on eyewitness accuracy studies criminal identification:

Can a witness correctly recognize a suspect?

What effect does a lineup have?

How does feedback affect confidence?

Those experiments provide important knowledge about human perception and memory. The National Academies concluded that these cognitive limits can have substantial consequences for eyewitness identification.

But historians frequently ask different questions.

They may want to know:

How did civilians experience occupation?

What did soldiers believe about their commanders?

How was racial prejudice expressed?

What did people fear during an epidemic?

These questions do not necessarily require witnesses to remember someone’s face with forensic precision.

A testimony containing several incorrect dates can still be excellent evidence about attitudes or remembered experience.

Psychological research therefore informs historical source criticism.

It does not provide a simple numerical reliability score for every memoir, diary, or oral history.

The Most Important Question: Reliable for What?

Consider a veteran who writes in 1980:

Every soldier in my unit hated our commander.

Is the statement reliable?

For proving that every soldier hated the commander?

Probably not.

One person cannot directly establish everyone’s thoughts.

For showing that the veteran personally disliked the commander?

Much stronger.

For showing how the veteran remembered unit morale decades later?

Excellent.

For establishing that dissatisfaction existed?

Potentially useful, especially if other independent evidence agrees.

The question:

Is this reliable?

is too broad.

Replace it with:

What specific historical claim can this testimony reasonably support?

A Historian’s Reliability Test for Eyewitness Accounts

When evaluating eyewitness testimony, work through these questions.

1. Who is the witness?

Identify the person’s background and relationship to the event.

Were they:

a participant,

observer,

official,

journalist,

civilian,

soldier,

victim,

political supporter,

opponent?

2. Where were they?

Determine physical proximity.

Could they actually see what they describe?

3. When was the account created?

Distinguish:

event → immediate account

from

event → decades → recollection.

The interval matters.

4. What was observed directly?

Separate:

I saw…

from:

Someone told me…

and:

We later learned…

These are different kinds of evidence.

5. What were the viewing conditions?

Consider:

distance,
lighting,
duration,
obstruction,
noise,
stress,
distraction.

6. What was the audience?

Was the account:

private,
public,
official,
confidential,
written for family,
produced for publication?

7. What was the purpose?

Did the witness want to:

inform,
persuade,
justify,
entertain,
defend,
accuse,
reassure?

8. What happened between the event and the account?

Did the witness encounter:

newspapers,
other witnesses,
documentaries,
official investigations,
books,
commemorations?

Later information may matter.

9. What does independent evidence say?

Compare the account with:

other eyewitnesses,
diaries,
letters,
photographs,
maps,
government records,
statistics,
physical evidence,
newspapers.

10. What can the testimony actually prove?

Keep the historical claim proportional to the evidence.

This final step is the most important.

A Practical Reliability Scale

Rather than labeling eyewitness testimony simply reliable or unreliable, it can be useful to think in terms of evidence strength.

Stronger evidence

An account may deserve greater weight when it was:

recorded close to the event,
based on direct observation,
produced under good viewing conditions,
specific about what the witness could and could not see,
independent of other accounts,
corroborated by separate evidence.

Requires greater caution

More caution may be appropriate when:

the account was produced decades later,
the witness describes events far outside their location,
important claims came from hearsay,
the witness encountered extensive later information,
the account serves strong reputational or political interests,
independent evidence contradicts it.

These are not mechanical rules.

They are prompts for investigation.

Example: A Battlefield Eyewitness

Imagine a soldier writes:

By noon the enemy was completely defeated. I watched thousands of them flee.

The letter was written that evening.

Strength

The account is contemporary.

Problem

The soldier occupied one small section of the battlefield.

Historical conclusion

The source provides good evidence that the soldier perceived a major enemy retreat from that position.

It does not independently establish that the entire enemy army was defeated.

A historian would compare the letter with:

maps,
other units’ records,
opposing army reports,
casualty information,
subsequent movements.

The testimony is useful precisely when its limits are understood.

Example: A Memoir Written Fifty Years Later

Imagine the same soldier writes fifty years later:

We knew immediately that the battle had changed the course of the war.

This should raise a different question.

Did soldiers actually understand the battle that way at the time?

Perhaps.

But later historical knowledge may have influenced the memory.

If contemporary letters say:

I have no idea what today’s fighting accomplished,

the contrast becomes significant.

The memoir is still valuable.

It may tell us how the soldier later incorporated the battle into the story of the war.

Memory itself becomes historical evidence.

Eyewitness Accounts Can Be Wrong and Still Be Valuable

This may be the most counterintuitive lesson.

Suppose thousands of people falsely believed a rumor that their city would be invaded.

Historians establish that no invasion had actually been planned.

Does that make diary entries describing fear useless?

No.

If your question is:

Was an invasion planned?

the rumor is inaccurate.

If your question is:

Why did civilians panic?

the rumor may be essential evidence.

Historical significance does not depend entirely on factual accuracy.

People act on what they believe.

False beliefs can produce very real historical consequences.

Frequently Asked Questions

Are eyewitness accounts reliable historical sources?

They can be, but reliability varies by claim and circumstance. Perception and memory have known limitations, and eyewitness accuracy can be influenced by viewing conditions, stress, delay, suggestion, and later information. Historians therefore contextualize and corroborate eyewitness testimony rather than accepting it automatically.

Why can eyewitness testimony be inaccurate?

A witness may have poor visibility, limited attention, incomplete information, high stress, fading memory, or exposure to misleading information after an event. Modern memory research shows that recollections can change during encoding, storage, and retrieval.

Can a witness sincerely remember something that did not happen?

Yes. Memory errors do not require dishonesty. Experimental research on misinformation shows that misleading post-event information can alter later memory reports even when participants are attempting to answer accurately.

Does stress make eyewitness memory less reliable?

High stress can impair aspects of eyewitness memory. A major meta-analysis found negative effects of heightened stress on both identification accuracy and recall. However, the relationship is more complicated than saying that every stressful memory is unreliable.

Does eyewitness memory get worse with time?

Some information is lost or becomes more difficult to retrieve with delay, and longer intervals can also increase opportunities for post-event information to influence recollection. Research has found reduced accuracy and greater vulnerability to misinformation under some delayed conditions.

Are confident eyewitnesses more accurate?

Sometimes, under particular conditions. Research indicates that an eyewitness’s initial high confidence can meaningfully predict identification accuracy when the first memory test is fair and uncontaminated. Later confidence is harder to interpret because feedback and other experiences can change it.

Can witnesses influence each other’s memories?

Yes. Research on memory conformity has shown that information supplied by another witness can influence subsequent recollection. Historians therefore need to determine whether apparently corroborating accounts are truly independent.

Are contemporary eyewitness accounts more reliable than memoirs?

Not always. A contemporary account may preserve immediate perception but incomplete information. A memoir may contain greater hindsight but weaker memory for details. Each can answer different historical questions.

Is a diary more reliable than oral history?

Neither format is automatically superior. Diaries are often closer in time to events, while oral histories may allow witnesses to explain experiences in greater depth. Both require contextual analysis and corroboration.

Can an inaccurate eyewitness account still be historically useful?

Absolutely. It can reveal what a person believed, feared, expected, remembered, or wanted others to believe. Those attitudes may themselves help explain historical behavior.

How do historians verify eyewitness accounts?

They compare them with independent testimony and other evidence such as letters, diaries, official records, photographs, newspapers, maps, statistics, physical evidence, and later scholarship. Corroboration is a core part of historical reasoning.

So, Should We Trust Eyewitnesses?

The wrong choices are:

Believe eyewitnesses because they were there.

or:

Ignore eyewitnesses because memory is imperfect.

Neither is good historical method.

Instead:

Listen carefully.

Identify what the witness actually experienced.

Determine when the account was created.

Separate observation from hearsay.

Consider memory, motive, audience, and later influence.

Compare the account with independent evidence.

Then decide which conclusions the testimony can support.

This is how eyewitness evidence becomes historical evidence.

Final Takeaway

Eyewitness accounts are neither perfect windows onto the past nor hopelessly unreliable stories.

They are something more interesting:

human evidence.

A witness sees from one position.

Not every position.

A witness remembers some details.

Not every detail.

A witness interprets what happened using the information available at that moment.

Later, memory may change as new information, conversations, hindsight, and repeated storytelling reshape the experience.

Modern research confirms that human perception and memory have important limitations and can be affected by stress, delay, suggestion, and later information.

But eyewitness testimony remains indispensable precisely because historical questions extend beyond the simple reconstruction of objective facts.

Historians also want to know:

What did people believe was happening?

What could they see?

What frightened them?

What did they misunderstand?

How did different participants experience the same event?

How did they remember it afterward?

For those questions, eyewitness accounts can be extraordinarily revealing.

The historian’s responsibility is not to choose between trust and distrust.

It is to determine how far the evidence allows us to go.

A single eyewitness rarely gives us the whole past.

But placed beside other witnesses, documents, photographs, objects, statistics, and contextual evidence, that testimony can help reconstruct something much richer:

not only what happened,

but how what happened was seen, experienced, understood, and remembered by the people who were there.