Thinking Like a HistorianChapter 1
Reading a Source
In this chapter
This chapter gives you the moves a historian makes before believing a document, and the checkable core of each — the part you can name and be right or wrong about. The moves are , , , , the guard against in your own reading, and . Each is practiced here on documents you can check for yourself.
Primary or Secondary Depends on the Question You Ask
A primary source was made at the time under study, or by someone who took part or watched; a secondary source was made afterwards, from other sources, to describe or interpret what happened.
Ask when and by whom the source was made, relative to the events — then ask what your question is. The same object can be primary for one question and secondary for another. Name the type for the question, not for the object.
A newspaper front page from the morning after an election (hypothetical) is primary for what readers were told that morning. A historian's book written fifty years later from such papers is secondary — until your question becomes how historians of that later decade wrote about the vote, when the same book turns primary.
"Primary means an original old document." Age and format are beside the point: yesterday's message is primary for yesterday; a printed edition of a medieval chronicle stays primary for medieval events. "Primary sources are more reliable — the writer was there." Closer is not truer: a witness saw part, remembered some and wrote for someone, which is what examines next. "Secondary sources are just opinion." A careful one shows the evidence behind each claim; the label marks distance, not quality.
Sourcing: Who Made It, When, for Whom, and Why
Sourcing is reading a document's origin before its content — who made it, when, for whom, and why — to fix what the maker could and could not have known, and had reason to say or leave out.
Four questions, each yielding something checkable. When: anything after the date the maker could not know, and the longer the gap to the writing, the more memory has done to the record. Who: what could this person see and hear, and what came second-hand. For whom, and why: what would this maker have reason to stress, soften or omit. A statement about something its author could not have known counts for nothing.
A general's memoir (hypothetical), written forty years after a battle he lost, for a public that blamed him — worked through below.
"Read the content first; who wrote it is background." Sourcing comes first because it decides what the content can be evidence of. "An eyewitness tells you what happened." A witness saw part, from one place; even one diarist's entry mixes what he saw with what he was told. "A writer with a motive is worthless." A motive tells you where to press, not to discard.
The general's memoir, sourced before it is read: 1. When. Forty years after the battle: everything he "remembers" has passed through four decades, and confident detail after that long is not proof of accuracy. Anything he says about what was known "at the time" must be checked against dated documents. 2. Who. A commander, present, with one field of view: strong on his own orders and what reached his tent, weak on what happened at the far flank until it was reported. 3. For whom, and why. Readers who blamed him; a defense. His account of his own decisions is where to press hardest; his account of the weather and the ground, which he had no reason to shade, is where to relax. 4. Then the content. Each claim is now labeled: seen or told, before or after the outcome was known, interested or indifferent — and read accordingly.
Thucydides, on how he came to know what he wrote, in the fifth century BCE:
> "And with reference to the narrative of events, far from permitting myself to derive it > from the first source that came to hand, I did not even trust my own impressions, but it > rests partly on what I saw myself, partly on what others saw for me, the accuracy of the > report being always tried by the most severe and detailed tests possible. My conclusions > have cost me some labour from the want of coincidence between accounts of the same > occurrences by different eye-witnesses, arising sometimes from imperfect memory, > sometimes from undue partiality for one side or the other. ... In fine, I have written > my work, not as an essay which is to win the applause of the moment, but as a possession > for all time."
Sourced in his own words: access (what he saw, what others saw for him), method (reports tested; witnesses differing by memory and by side), and purpose (permanence, not applause). Each is a claim about the author, and each can be checked against the rest of his book.
The Titanic struck an iceberg on the night of April 14, 1912, and sank in the early hours of April 15; more than 1,300 people died. Newspapers dated April 15 could not yet know this. The Washington Times' last edition that day announced the liner "kept afloat by water-tight compartments," being towed into Halifax; Honolulu's Evening Bulletin printed "All Titanic Passengers Saved". On April 16 the New-York Tribune's front page read "1,340 Perish as Titanic Sinks". The date fixes what each could know: sourced, the April 15 pages are evidence of what had reached shore that day, not of what happened at sea.
Contextualization: What Was Normal, Known, and Meant at the Time
Contextualization places a document in its own time and place — what was happening, known, normal, and what its words meant then — so that a phrase, a silence or an act is read as its first readers read it.
Fix three things before trusting your reading. The words: each period has "a language of its own," and a word's present sense is the last thing to assume. The moment: the events, calendar and common knowledge of the date. The genre: a silence proves absence only in a document that would naturally have recorded the thing.
A London letter (hypothetical) dated "the 20th of February, 1660." Until 1752 the English legal year began on March 25, so by that reckoning the letter belongs to what we now call February 1661. Nothing in the document changed; what the reader brought to it did.
"Old words mean what they mean now." Meaning drifts by period, place and author: vel means "and" and suffragium "help" in medieval Latin. "If it doesn't mention it, it wasn't happening." Most of what happens is never written down, and most of what was written is lost. "Context is background color." Context is what the words meant to those who wrote and read them; without it the reading is yours.
Corroboration: What a Second Document Confirms, and What It Does Not
Corroboration checks a claim in one document against an independent second document and reports, claim by claim, what the second confirms, what it contradicts, and what it leaves untouched.
Split the first document into separate claims — place, time, actors, weather — and test each; corroboration is never of the document as a whole. For each claim the second document asserts it, denies it, or is silent, and silence is neither. Count agreement only between independent observations — not one document copying another, nor two writers repeating a story that was going round; two independent witnesses to the same specific fact is the working standard. Where they differ, do not split the difference; return to .
Two diaries of one September night, in the passage below, sorted claim by claim in the steps that follow it.
"Two documents agree, so it's true." Only if they are independent and agree on the specific claim: Fling's two memoirists at Versailles told the same tennis-court story years later, when the tale was already current — agreement that established nothing. "They disagree, so one is lying." Position and hearsay explain differences before dishonesty does. "The second doesn't mention it, so it contradicts it." Silence leaves the claim resting on one witness — neither more nor less.
Two diaries, one fire. Samuel Pepys, at home near the Tower of London, Sunday, September 2, 1666:
> "Jane called us up about three in the morning, to tell us of a great fire they saw in the > City. ... By and by Jane comes and tells me that she hears that above 300 houses have > been burned down to-night by the fire we saw, and that it is now burning down all > Fish-street, by London Bridge. ... [The Lieutenant of the Tower] tells me that it begun > this morning in the King's baker's house in Pudding-lane ..."
Later that morning, from a boat on the river:
> "... and the wind mighty high and driving it into the City; and every thing, after so > long a drought, proving combustible."
John Evelyn, who went to the Bankside in Southwark to watch across the river, under September 2 and 3, 1666:
> "This fatal night, about ten, began the deplorable fire, near Fish street, in London. > ... The fire having continued all this night ... when conspiring with a fierce eastern > wind in a very dry season, I went on foot to the same place; and saw the whole south part > of the city burning."
Pepys's claims, one at a time, against Evelyn: 1. Where it began. Pepys: Pudding Lane, by Fish Street, second-hand from the Lieutenant of the Tower. Evelyn: "near Fish street". Confirmed at the level of the street; the baker's house rests on Pepys alone. 2. The wind. "Mighty high and driving it into the City" against "a fierce eastern wind". Confirmed, by two men writing separately and in different words. 3. The dry spell. "So long a drought" against "a very dry season". Confirmed. 4. The hour. Pepys was woken at about three by a fire already burning, and was told it "begun this morning"; Evelyn writes "about ten". Different — and not yet a lie: Pepys was near and awake at three, writing down what he saw and was told that day; Evelyn watched from across the river. decides which account to prefer, and here it prefers Pepys's. 5. What Evelyn never mentions. The baker's house, Jane's report of houses burned. Silence: those claims are neither confirmed nor contradicted, and still rest on one diary.
Presentism: Reading the Past With Knowledge It Did Not Have
Presentism is reading the past with knowledge, words or standards its people did not have: judging a choice by an outcome the actor could not see, taking a word in its present sense, or assuming they knew what we know.
Point to the sentence and ask what it depends on. Something learned after the actor acted — hindsight, not history. A word carrying its present meaning — is the check. A present-day category or standard treated as theirs — name it. Then say what they knew and meant, on their terms, and keep what you know in a separate sentence marked as yours.
A reader (hypothetical) says a medieval chronicler "lied" in reporting that a comet foretold the king's death. The chronicler shared his age's belief in omens; calling the report a lie imports our astronomy into his sentence. The checkable error is the imported knowledge.
"They should have seen it coming." An outcome known to you was not a sign to them. "Old text, same word, same meaning." Meaning is fixed by period, place and author. "Avoiding presentism means never calling anything in the past wrong." Whether to judge the past by present values is a live disagreement (below); the reading error is not, and judgment comes after an accurate reading.
After Is Not Because: Causation and Sequence
A sequence says one event followed another; a causal claim says the second happened because of the first — a stronger claim that needs more than order in time.
Beyond order in time, "because" needs three things. A link you can describe: how the first produced the second. Reason to think the second would have gone differently without the first — so look for other changes in the same season, since a common cause or a coincidence explains many sequences. And a name for the kind of cause: the trigger that set things off, or an underlying condition that made the outcome possible. "The" cause is rarely an honest answer.
A town chronicle (hypothetical) records that a bridge opened in spring and the market across the river doubled by summer. That is sequence; the steps below show what "because" would take.
"It came right after, so that caused it." Post hoc, ergo propter hoc — and "a hundred other antecedents" can claim the same title. "Every event has one cause, the thing just before it." Fling names lack of bread as the immediate cause of the Paris rising of October 1789 and traces the political series beneath it. "Telling it in order is explaining it." Chronological order and causal connection are different achievements.
From "after" to "because," for the bridge and the market: 1. State the sequence and nothing more. The bridge opened in spring; the market had doubled by summer. This is all the chronicle supports. 2. Name the link. Traders and buyers from the far bank could now cross cheaply; the growth would show first in goods and faces from across the river. 3. Look for rivals in the same season. A good harvest, a new tax on the old market, a fair moved to the far bank — each would double a market without any bridge, and any of them present weakens the claim. 4. Name the kind of cause. If the link holds and the rivals fail, the bridge is a trigger; the far bank's growing population, if it was growing, is the condition that made a bridge matter. Say which — never "the cause."
Reading One Source End to EndSynthesis
Take one document — a minister's memoir of the government he served, written thirty years after it fell, for readers who blamed him. For your question about the fall it is a , because he took part — and primary again for how the fall was later remembered. Then : thirty years of memory stand between him and the events, his position opened some rooms and not others, and his purpose is defense — press hardest where the memoir clears him. next: each phrase carries the meanings of its decade, and a silence about a scandal counts only if a memoir of that kind would have mentioned it. comes claim by claim against an independent record — a diary kept at the time — and says which claims it confirms, which it contradicts, and which it never touches. Read your own reading for : what he "should have foreseen" is only what was knowable then. And when he says the fall came because of one decision, ask whether he has shown a link and ruled out rivals, or only that the decision came first — must be earned; sequence is all a memoir gives for free.
Sources
- 1 Library of Congress, Getting Started with Primary Sources — definitions of primary and secondary sources and the sourcing and context questions (US government work, public domain) — source
- 2 Library of Congress, Teacher's Guide: Analyzing Primary Sources — Observe / Reflect / Question (US government work, public domain) — source
- 3 Langlois and Seignobos, Introduction to the Study of History (1898), trans. Berry, Book II ch. VI §III — the language of the time, the country and the author; the rule of context — source
- 4 Langlois and Seignobos, Introduction to the Study of History (1898), Book II ch. VII §§IV–V — reasons for doubting good faith (the author's interest, official documents, 'authentic' is not 'sincere') and accuracy (the author badly situated for observing) — source
- 5 Langlois and Seignobos, Introduction to the Study of History (1898), Book II ch. VIII §§II–V — a statement about a fact the author could not know is null; contradictions are not to be split; agreement counts only between independent observations, and proves more when confined to a few circumstances — source
- 6 Langlois and Seignobos, Introduction to the Study of History (1898), Book III ch. I §IV — the historian imagines the past on the model of the present, 'one of the principal sources of error in history' — source
- 7 Langlois and Seignobos, Introduction to the Study of History (1898), Book III ch. III §II — the argument from silence and the two conditions under which it is admissible — source
- 8 Fling, The Writing of History (Yale University Press, 1920), ch. I–II, pp. 18–19 and 43–45 — remains versus tradition; 'at least one human brain stands between us and the fact' — source
- 9 Fling, The Writing of History (1920), ch. IV, pp. 69–71 and 78–80 — the time of writing and memory; dating by termini; separating first-hand from second-hand information by what the witness could have seen — source
- 10 Fling, The Writing of History (1920), ch. V, pp. 96–99 — dependence between sources; the tennis-court story told by two apparently independent memoirists — source
- 11 Fling, The Writing of History (1920), ch. VI, pp. 103–105 — an affirmation is not a fact; two independent witnesses to the same detailed fact as the condition of certainty — source
- 12 Fling, The Writing of History (1920), ch. VII, pp. 135–136 and 143–145 — 'facts and facts do not constitute history': chronological order versus causal connection; the immediate cause of the Paris insurrection of October 5, 1789 (the lack of bread) and the political series beneath it — source
- 13 J. S. Mill, A System of Logic, 8th ed. (1882), Book V ch. V §§4–5 — reading history through the men and things of one's own age; post hoc ergo propter hoc, and 'a hundred other antecedents'; and Book III ch. VIII §6, the fifth canon — two things that vary together may be 'two different effects of a common cause' — source
- 14 Thucydides, History of the Peloponnesian War, Book I ch. 22, trans. Richard Crawley (1874), Project Gutenberg #7142 (public domain) — source
- 15 The Diary of Samuel Pepys, entry for September 2, 1666 (Lord's day), ed. Henry B. Wheatley (1893), Project Gutenberg #4200 (public domain) — source
- 16 The Diary of Samuel Pepys, ed. Wheatley (1893), editorial note at January 1659–60 on the English legal year beginning March 25 until 1752 (public domain) — source
- 17 The Diary of John Evelyn, entries for September 2–4, 1666, ed. William Bray, vol. 2 (M. Walter Dunne, 1901), Project Gutenberg #42081 (public domain) — source
- 18 Calendar (New Style) Act 1750 (24 Geo. 2 c. 23), s. 1 — 'the said supputation, according to which the year of our Lord beginneth on the twenty-fifth day of March, shall not be made use of from and after the last day of December one thousand seven hundred and fifty-one' — source
- 19 Library of Congress, Sinking of the Titanic: Topics in Chronicling America — timeline (iceberg struck April 14, 1912; sank in the early hours of April 15; over 1,300 dead; conflicting reports the following morning; sinking confirmed April 16) — source
- 20 The Washington Times (Washington, DC), April 15, 1912, Last Edition, p. 1: 'Liner Titanic Kept Afloat by Water-tight Compartments Being Towed into Halifax, N.S.' (Chronicling America, public domain) — source
- 21 Evening Bulletin (Honolulu, HI), April 15, 1912, 3:30 Edition, p. 1: 'All Titanic Passengers Saved' (Chronicling America, public domain) — source
- 22 New-York Tribune (New York, NY), April 16, 1912, p. 1: '1,340 Perish as Titanic Sinks; Only 886, Mostly Women and Children, Rescued' (Chronicling America, public domain) — source
- 23 Lynn Hunt, 'Against Presentism', Perspectives on History (American Historical Association), May 2002 — cited as a position, not adapted — source
- 24 David Armitage, 'In Defense of Presentism', in Darrin M. McMahon (ed.), History and Human Flourishing (Oxford University Press, 2023); author's preprint — cited as a position, not adapted — source
- 25 The Anglo-Saxon Chronicle, entry for A.D. 1066 — an eleventh-century chronicle of the Conquest year (Harold, the comet, William's landing), trans. James Ingram (1823) with readings from J. A. Giles (1847), Project Gutenberg #657 (public domain); the ladder's 1066 and its 'chronicles from the eleventh century' rest on it — source
- 26 The Diary of Samuel Pepys, ed. Wheatley (1893), entries for May 8–9, 1660 — the King 'proclaimed yesterday with great pomp' in London (May 8, 1660): the ladder's date line for a king restored by February 1661 (public domain) — source
- 27 London Museum, 'The Great Plague of 1665' — bonfires 'to fend off miasmas, the bad air blamed for spreading the plague'; the disease is caused by Yersinia pestis and spread by the bites of infected rat fleas — consulted for the presentism ladder's facts (D-070; nothing adapted) — source
- 28 Institut Pasteur, 'Alexandre Yersin, the man who discovered the bacterium responsible for the plague' (2017) — the plague bacillus isolated in Hong Kong on June 20, 1894; consulted for the presentism ladder's date (nothing adapted) — source
- 29 OpenStax, World History, Volume 1: to 1500, §1.2 Primary Sources — consulted for scope and content corroboration only; CC BY-NC-SA 4.0 as displayed 2026-08-18, so nothing is adapted from it — source
- 30 OpenStax, World History, Volume 1: to 1500, §1.3 Causation and Interpretation in History — consulted for scope corroboration only (levels of causation; 'historical empathy'); CC BY-NC-SA 4.0 as displayed 2026-08-18, so nothing is adapted from it — source