E. H. Carr Writing Style

Carr opens What Is History? by putting two prefaces to the same multi-volume history side by side, Acton's confident one and George Clark's much warier one, then refusing to referee them. The instruction arrives underneath. Study the historian before you study his facts. That reordering of attention runs the whole book, and it is a writing decision before it is a philosophical one, because it settles what belongs in the opening paragraph and what has to wait until the reader has been made suspicious.

His paragraphs move by pressure. A claim goes down flat, in the plainest words available, and then a limit arrives, and the limit is the half worth studying, because it names the conditions under which the flat claim would stop holding. Watch him work the word fact. He lets you keep the ordinary meaning for a while, then produces the gingerbread seller kicked to death by a crowd at Stalybridge Wakes, a thing that certainly happened and that historians had left alone, and asks whether it counts as a fact of history yet.

Imitating the sentences gets a writer nowhere, because the sentences come last. The order underneath them does the work: set up the tempting version, let the reader hold it, then bring the case that breaks it. A reader can take in a denial only after the thing denied has been made plausible, which is why Carr spends real space making the comfortable view attractive before he removes it. Skip the setup and identical sentences land as scolding.

Two books show what he chose not to do. The Romantic Exiles follows Herzen and his circle through their quarrels and love affairs in exile, and it reads like a novel. The Twenty Years' Crisis argues, and it argues by running utopia against reality for most of its length before refusing to leave the reader at either pole. Carr could write scene. He mostly declined. The decision is worth copying if your drafts keep reaching for colour where the paragraph needs a condition instead.

Definition before use

The working meaning of a key word goes in one line, at the moment the word starts to carry load, and it arrives with what it excludes. That heads off the argument where two sides agree in words and disagree in meaning, which is where most disputes about progress or objectivity end up. The reader feels the floor firm up. The cost is commitment for the rest of the piece, and a definition you cross later is worse than one you left out, so the tool works only in company with the claim-first shape it supports.

The sorted cause

Causes get split by usefulness rather than by force. Which of these would matter in a case like this one, and which explains the case in front of us and nothing else? Carr runs the sort on Robinson's death in What Is History?, and the reader is left with a short list that travels and a longer one that stays put. The effect is relief, because a page of causes in no particular order exhausts anyone. The difficulty is honesty about the residue, since the cause you set aside was a cause, and pretending otherwise is how a tidy argument turns false.

The opponent quoted whole

Opposition appears in its own words, at a length that lets it look reasonable, and the reply concedes a boundary before it redirects. Readers trust this fast. They can see that the hard part was left in, and the trust carries over to claims they have no way of checking. The hard part for the writer is understanding the other side well enough to state it cleanly, then resisting the pleasure of mocking it. It pairs with definition before use, since the boundary in the reply is often a definitional limit rather than a retreat.

The conceded inch

A small point goes to the other side, chosen because it costs little and buys scope. The concession shrinks what the essay claims, which is what stops the claim sounding absolute, and it does that without a paragraph of hedging around it. Trust follows from the display of discrimination. He can tell a strong objection from a weak one, so his refusals mean something when they come. Make the inch too generous and the thesis leaks. Make it too grudging and the whole move reads as a manoeuvre.

Selection made visible

Carr shows the reader the sieve. Where the facts came from, who kept them, what fell out along the way, with the Stresemann papers as his standing example, since the surviving portrait had been filtered by a secretary and then filtered again by a translator. Doing this in your own work solves a credibility problem cheaply, because a reader who can see your criteria stops wondering what you left out. It constrains you too. A sieve you have shown is a sieve you have to keep using for the rest of the piece.

Strength matched to evidence

Some sentences land as conclusions and some stay conditional, and the difference tracks the evidence rather than the mood of the paragraph. That fixes the credibility problem of a writer who sounds equally sure about everything, where a reader has no way to separate the load-bearing claims from the decoration. The effect is safety. Someone is exercising judgment on your behalf. The work behind it is a constant audit of your own sentences, asking which is observation, which is inference and which is interpretation, and it holds the rest of the toolkit together, because a chain of consequences built out of overstated links comes apart at the first one a reader leans on.

Antithesis that sorts

Carr sets two positions against each other to sort them, not to stage a fight. The Twenty Years' Crisis runs on the pair utopia and reality and keeps returning to it, and each return leaves a sharper sense of what either pole can deliver and what it cannot. Then he declines the tidy ending. Neither pole survives alone, and the refusal to settle is the argument the book was built to make. Copied badly, the device turns into a see-saw of on the one hand and on the other. Carr's poles have consequences attached, so choosing one commits you to something a reader can hold you to later.

The objection brought forward

He answers the reader's rebuttal before the reader has finished forming it. Timing is the trick. Resistance that goes unaddressed hardens into distrust within about a paragraph, and after that no amount of reasoning gets through, so the caveat arrives early, where it sharpens the claim instead of apologising for it. There is a limit to this. Anticipate everything and the argument stops moving, anticipate only the easy ones and the reader notices which objection you walked around.

The question as a hinge

Look at the title he chose. A question, and a plain one, and the chapters keep answering it in ways that refuse to be final. Inside a paragraph the same device closes one unit of reasoning and opens the next, which does the job a signposting sentence would do without the throat-clearing that comes with it. A question also smuggles in criteria, since asking which causes are worth stating tells the reader that usefulness is now the standard. Used carelessly it becomes filler. A reader who senses that no answer is coming stops taking the next question seriously.

Concession with a boundary

He gives ground on purpose, then fences what he gave. The concession lowers the reader's guard and the fence does the arguing, because it says what the concession costs and where it stops. That is how he moves between stages of an argument without a jolt: from yes, to yes under these conditions, to a claim that has survived contact with the strongest thing said against it. The failure mode is a soft fence. Concede without a boundary and the paragraph reads as surrender, which is the mushy centrism people mistake for balance.

Copying the verdict sentences and skipping the build

Carr's flat declaratives are the last brick, and imitators start with them. The prose then promises a structure that is not in the room, and readers feel the gap before they can name it, which produces suspicion where the writer wanted assent. Look at what precedes a firm sentence in What Is History?. A premise has been installed, a term has been fixed, an objection has been given space to sound reasonable, and only then does the short verdict arrive. Build the courtroom before you deliver the verdict, or the verdict is a man shouting in an empty room.

Reading selection as a licence to stop arguing

The idea that a historian selects gets picked up as permission to stop making a case, and the result is fog. Pages of commentary about perspective, with nothing in them a reader could check. Carr uses selection the other way round. Once you admit that you chose, you owe the criteria, and a criterion can be tested where a confession of bias cannot. Writers who miss this sound evasive while believing they sound sophisticated, and readers leave in the space between the two. The fix is small and unpleasant. Name the standard you selected by, in the text, where you can be held to it.

Stacking abstract nouns until the page floats

Process, structure, agency, context, framework. Carr's subject matter invites them and imitators lay them end to end. Steven Pinker calls this family metaconcepts, and the trouble is that a sentence built out of them cannot be pictured or tested, so a reader has no way of telling whether it claims anything at all. Carr keeps his abstractions tethered to somebody deciding under pressure. His fish are being caught by a historian who chose the water and chose the tackle. Run the check on your own paragraph. An abstract noun that changes nothing about what the reader expects next is packing material.

Staging objections you can knock over

A token line of dissent followed by an easy reply looks like balance and reads like theatre. The reader hears the difference, because a real objection forces the writer to give something up and a decorative one leaves the argument where it already was. Carr's objections cost him. They narrow his scope, move his definitions, and put limits into the main text where a lesser writer would have used a footnote. Here is the test before you keep one: if answering the objection changed nothing in the paragraph above it, the objection is set dressing.