Reading Critically and Analyzing Rhetoric

A text is a machine somebody built to do something to you. Chapter 1 asked you to think about the situation you write into; this chapter turns the lens around and asks you to take apart the situations other people wrote into. That is a practical skill, not a literary parlor trick. The discharge instruction patients ignore, the policy memo that quietly shifts blame, the op-ed that makes you nod before you have checked a single fact — all of them are engineered, and the engineering becomes visible once you know where to look. You will learn to read in passes instead of one anxious sweep, to separate what a paragraph says from what it does, to name the appeals a writer is running and the assumption holding the whole thing up, and finally to write an analysis that explains how a text works rather than retelling what it said. Every move here comes with a worked example attached.

Section 1Two Kinds of Reading

What it says versus what it does

Most of the reading you do is extraction. You want the dose, the deadline, the due date, the ending. That is reading for information, and it is the right tool most of the time. Rhetorical reading asks a second question on top of the first: not only what is this telling me but what is this doing to me, and by what means. The second question is where analysis lives, and it is invisible until you deliberately ask it.

Take one real-feeling sentence from a hospital-wide email. (Every sample passage in this chapter was written for teaching; none is a quotation from a real publication, though each is modeled on texts you will actually meet.)

Sample: “Due to ongoing staffing challenges, weekend shift differentials will be adjusted effective March 1.”

What it says: weekend pay changes on March 1. What it does is a longer list. It puts the excuse before the news, so you meet the reason while you are still neutral. It uses an agentless passive — differentials “will be adjusted” — so that no human being appears anywhere in the sentence doing the adjusting. It selects “adjusted,” which could mean either direction, over “cut,” which could not, and “challenges” over “shortage,” which would raise the question of who is responsible for the shortage. Four decisions, one sentence, twelve seconds of your attention.

Notice that the “does” answers all take the form of active verbs with the text as the subject. That is the habit to build. A text concedes, reframes, pre-empts, narrows, escalates, defers, subordinates, universalizes, changes the subject. If your note says “talks about pay,” you have written a summary. If it says “buries the agent so no one can be argued with,” you have written analysis.

A three-pass protocol

Trying to comprehend and evaluate at the same time is why hard reading feels like drowning. Your working memory cannot hold an unfamiliar argument in place while also auditing it. Split the job into passes and each pass gets easy.

The protocol also gives you permission to be confused on the first pass, which is the real reason students abandon difficult texts. Confusion at minute three is expected. Confusion after pass three is information: it usually means the writer skipped a step, or you are missing a term the writer assumed you had.

Annotation that earns its ink

Use a small, fixed code so that annotation costs you nothing to decide. Nine symbols are plenty: C for the claim, R for a reason, E for evidence, W for a warrant or buried assumption, CA for a counterargument the writer raises, DEF where a key term gets defined, MOVE for a rhetorical maneuver worth naming, !! for the strongest moment in the piece, and a bare ? wherever a sentence does not follow from the one before it.

Here is a marked passage. The bracketed notes are what the margin should contain.

Sample, annotated: “Hospital administrators are not villains. [CA + MOVE: concedes early to buy fairness credit] They are solving a real budget problem with real constraints. [MOVE: grants the opposition its best case] But a budget is a statement of priorities, not a law of physics. [C: this is the actual claim, dressed as an aphorism] Last year this system spent more on its executive search firm than on the four night-shift positions it left unfilled. [E: one comparison, no source given — ?]”

Three things fall out of four sentences. The concession is doing structural work rather than intellectual work — the point conceded never constrains the conclusion. The claim arrives disguised as a truism, which is a way of making a contestable statement feel obvious. And the evidence is a single unsourced comparison chosen for its moral shape, not its magnitude. None of that is visible on a first pass, and all of it is visible in the margin.

Section 2Ethos, Pathos, Logos: Reading the Appeals

Aristotle divided persuasion into three pisteis, usually translated as the artistic proofs: appeals grounded in the speaker’s character (ethos), in the audience’s emotional state (pathos), and in the reasoning itself (logos). Translators and scholars still argue about the details — a common reading is that Aristotle meant ethos to be constructed within the speech rather than imported from the speaker’s prior reputation, which is not how most people use the word today. That disagreement is worth knowing, because it changes what you look for. If ethos is built in the text, you can point at the sentences that build it.

Ethos: why should I believe you?

Ethos is the answer a text gives to the question “why should I take your word for this?” It is assembled out of small moves: claiming expertise, claiming proximity to the events, demonstrating fairness by conceding something, signaling shared values, and simply writing in the register the audience expects. It is also destroyed by small moves — a sneer, a typo in a formal document, a certainty the evidence cannot support.

Sample A, the opening of a newspaper op-ed by a nurse we will call Alvarez: “Every nurse I know has a story about the night that nearly broke them. Mine was a Tuesday in February, when I was the only RN on a unit of thirty-two patients and a man in room 14 was quietly going into septic shock while I was two doors down, charting. He lived. I have not stopped thinking about the version where he didn’t. Legislators call staffing ratios a budget question. From where I stand, it is an arithmetic problem with a body attached.”

Count the ethos moves. “Every nurse I know” claims membership in a group before it claims anything about the world. “Mine was a Tuesday in February” claims proximity. The operational detail — the room number, the charting, the two doors of distance — claims a kind of reliability that no credential could: this is a person who remembers procedure under stress. And “He lived” is the subtlest move of the four, because admitting that no harm occurred forfeits the strongest emotional card in the deck. Forfeiting it is exactly what makes the writer trustworthy.

Pathos: what am I being made to feel?

Pathos has an undeserved reputation as the dishonest appeal. It is not. Putting an audience into the emotional state where an argument can land is a legitimate and often necessary part of persuasion — nobody acts on a fact they feel nothing about. The analytic question is never “is there emotion here.” It is: does the feeling track the situation, and is it standing in for a reason it cannot supply?

In Sample A the strongest emotional device is the one the writer refuses to use. She does not narrate the patient dying. She says “the version where he didn’t” and stops, which hands the imagining to you. A harm you constructed yourself is markedly harder to argue with than one you were shown, and it costs the writer nothing in credibility because she never asserted it.

Compare a version that overplays the same hand: “Every legislator who votes against ratios has blood on their hands, and they know it.” The feeling has stopped tracking anything. It asserts knowledge of other people’s minds, converts a policy disagreement into a moral crime, and gives any undecided reader an easy reason to stop reading. That sentence damages ethos and logos simultaneously, which is the usual price of unearned pathos.

Logos and kairos: the shape of the reasons, and the timing

Logos is the appeal of the reasoning — the reasons offered, the evidence behind them, and the structure connecting the two. Sample A’s logos is implicit and compressed into a single ratio — one nurse, thirty-two patients — plus the scene that ratio produces: a deteriorating man two doors away. The final line makes the reasoning explicit by renaming the issue: not a budget question but “an arithmetic problem with a body attached.”

Sample B, the same claim in a research register: “In a retrospective cohort of 421 medical–surgical units, each additional patient per nurse was associated with a 7 percent increase in 30-day inpatient mortality after adjustment for case mix. These findings should be interpreted cautiously, as unmeasured differences in unit acuity may account for part of the observed effect.” Nothing here tries to move you, and the caution in the second sentence is doing real work: it is the writer telling you exactly how far the design lets the claim travel. (The underlying question is genuinely contested in the health-services literature; studies consistently find associations between nurse workload and outcomes, and researchers continue to disagree about how much of that association is causal.)

Kairos is the fourth term and the one students skip. It names the fitness of an utterance to its moment — the opportune time, the ripe occasion. A staffing op-ed published the week a ratio bill reaches the floor is a different act from the same words published in July. Kairos is closely tied to exigence, the problem or event that made the text necessary. When you cannot find the exigence, look at the date.

Section 3Finding the Argument Under the Prose

Claim, reasons, evidence

A claim is a statement a reasonable, informed person could disagree with. That is the whole test. “Nurse staffing is important” fails it — nobody is on the other side — so it is a topic wearing a claim’s clothing. A reason is what follows the word because. Evidence is whatever makes the reason believable to this particular audience: data, testimony, precedent, documents, examples.

Sample C, from a faculty-senate memo: “The college should stop requiring placement essays written in a single ninety-minute sitting. Timed writing measures typing speed and test nerves at least as much as it measures composing ability, and it samples one stage of a process that is by definition iterative. When our department piloted a portfolio option last spring, students placed into the same courses 84 percent of the time, and the students who moved, moved up.”

Pull it apart. Claim: drop the timed placement essay. Reason 1: the instrument measures the wrong construct. Reason 2: it samples a single stage of an iterative process. Evidence: a departmental pilot with 84 percent agreement and asymmetric disagreement. Now hold the evidence up to the light. If the two instruments agree 84 percent of the time, that is a decent argument for keeping the cheaper one. The entire case therefore rests on the 16 percent — on the students who moved, and on the claim that they moved up, meaning the timed essay had underplaced them. The memo never says this out loud. Finding the sentence the argument actually depends on, and noticing that the writer left it unwritten, is what analysis buys you.

The warrant: the part nobody writes down

Stephen Toulmin’s The Uses of Argument (1958) gave us the vocabulary most composition courses still use. His key contribution was the warrant: the general principle that licenses the move from evidence to claim. Warrants are usually unstated, because stating them feels insulting when they are shared — and because stating them is fatal when they are not. Toulmin added a qualifier (“probably,” “in most cases”) and a rebuttal (the conditions under which the claim would not hold), and reading for those two is a fast way to gauge how honest an argument is being with itself.

To surface a warrant, put the evidence and the claim side by side and finish this sentence: that only counts as a reason if you already believe…

• “We should hire more nurses because inpatient mortality falls.” — only counts if you believe lower mortality is worth this cost. Widely shared, so nobody writes it.
• “That source is unreliable because it was published on a blog.” — only counts if venue predicts reliability. Contestable, and increasingly weak.
• “Take the 8 a.m. section because that professor is tougher.” — only counts if harder is better for you. Very contestable, and it depends entirely on your semester.
• “Adopt the portfolio because it agrees with the timed essay 84 percent of the time.” — only counts if agreement with the old measure recommends the new one, which is exactly backwards.

Most arguments that feel slippery are slippery at the warrant. The claim is defensible, the evidence is real, and the invisible principle joining them is the thing you would never have signed if it had been printed on the page.

Reverse outlining a text you did not write

A reverse outline is a paragraph-by-paragraph inventory of an existing text: one column for what the paragraph says, one for what it does. You will use it on your own drafts in Chapter 3 and again as the main tool of global revision in Chapter 14. Used on someone else’s argument — here, the full op-ed that Sample A opens — it exposes structure that reading straight through will hide, because prose is designed to feel like a smooth surface.

The right-hand column is the essay’s architecture, and it is the raw material of a rhetorical analysis. Notice what the outline reveals that a summary would not: the claim is deliberately last, the concession is deliberately adjacent to it, and the research is doing borrowed work in the middle. You could summarize this piece perfectly and never notice any of that.

Section 4Reading the Sentence: Diction, Syntax, Framing

Diction and connotation

Denotation is what a word points at. Connotation is everything it drags along — the attitude, the register, the social world it belongs to. Two writers can report the identical fact and hand you two different arguments without stating either one. This is the most common place where persuasion happens invisibly, and it is easy to catch once you practice substitution: swap the word, see what changes.

When you find loaded diction, resist the urge to stop at “this is biased.” That is a verdict, not an analysis. The analytic sentence names the substitution and its consequence: calling the funding a handout does the argumentative work that the essay’s single paragraph of budget evidence cannot, because it settles the question of desert before any figures appear.

Syntax, emphasis and the end of the sentence

English loads meaning toward the end of a sentence. Readers hear the final position as the point — the stress position — and hear whatever is subordinated as background. So the same two facts, rearranged, become opposite arguments.

Two more structural habits are worth marking in the margin. Nominalization turns a verb into a noun and quietly evacuates the actor: “there was a failure of escalation” instead of “the charge nurse did not escalate.” You will meet nominalization again in Chapter 11 as a style problem; here it is a reading problem, because it is the standard way institutions describe their own mistakes. And sentence length is rhythm doing argumentative work. Sample A builds a thirty-seven-word sentence full of clinical detail and then lands two words: “He lived.” The short sentence feels true partly because it is short.

Framing, hedging and metadiscourse

Framing is the category an issue gets filed under, and it may be the highest-yield thing you can name in an analysis, because frames carry their own rules for settling disputes. A “budget question” is settled by weighing trade-offs; a “safety question” is settled by setting a minimum standard below which trade-offs are not allowed. Sample A’s closing line does not argue for the second frame. It simply performs it, and a reader who accepts the frame has conceded most of the argument without noticing.

Hedging (“may,” “appears to,” “one plausible explanation”) and boosting (“clearly,” “obviously,” “it is well established”) calibrate how much confidence a writer is asking you to extend. Opinion here genuinely splits: plain-style handbooks treat hedges as padding and tell you to cut them, while researchers on academic discourse — Ken Hyland’s work is the usual point of reference — argue that hedging does real epistemic and social work, matching claim to evidence and leaving the reader room to disagree. Both camps are partly right, and the workable test is simple: does the hedge match the evidence?

Finally, metadiscourse is the text talking about itself: “I will argue,” “having established this, I turn to,” “note that,” “the objection I take most seriously is.” It carries no content about the world, and it is enormously useful to a reader — and to an analyst, because a writer’s metadiscourse is their outline, published by accident. Underline every instance and you often have the reverse outline for free.

Section 5Writing the Rhetorical Analysis

The thesis that names a strategy

A rhetorical analysis argues about how a text works. Its thesis therefore has to make a contestable statement about the text’s machinery — the strategy, its effect, and the audience condition that makes the strategy pay or fail. If your thesis would still be true after you swapped in a different article, it is not a thesis. One scaffold, offered as training wheels rather than a template to submit: Facing [exigence] and readers likely to [resist in some specific way], the writer [strategy] in order to [effect] — a move that [succeeds / strains] because [reason].

The analytical paragraph: point, quotation, unpacking, consequence

Each body paragraph makes one claim about the text, quotes briefly, explains how the quoted language produces the effect, and connects that effect back to the essay’s larger argument. Chapter 9 covers the mechanics of quoting and Chapter 10 the citation formats; what matters here is the third step, which is the one students skip. Also retire two verbs. “Uses” and “says” describe nothing. Reach instead for concedes, reframes, qualifies, escalates, pre-empts, subordinates, universalizes, personalizes, withholds, defers, undercuts — each of these forces you to specify an action.

Draft (summary wearing an analysis costume): “Alvarez begins her essay by telling a story about a night when she was the only nurse on a unit with thirty-two patients. A man in room 14 went into septic shock while she was charting. She says he lived, but she still thinks about it. This story shows that nurse staffing is a serious problem and it makes the reader feel sad. Then she talks about the research.”

Revision: “Alvarez opens not with a proposal but with a debt. The near-miss in room 14 is told in operational detail — the room number, the charting, the two doors of distance — so that her credibility rests on procedural competence rather than on feeling, and the two-word sentence that resolves it (‘He lived’) refuses the catharsis the setup has promised. What the reader is left holding is the ending Alvarez declines to narrate. That withheld counterfactual does work no statistic can do: it makes the reader supply the harm, and a harm you imagined yourself is far harder to argue with than one you were shown. Only after the debt is incurred does the mortality research appear, arriving as confirmation of something the reader now feels they already knew.”

Four things changed. The subject of nearly every sentence became the text or the writer rather than the topic. Every claim about effect is tied to a specific piece of language. The paragraph names a structural choice (story before proposal) and explains what that ordering buys. And it ends on consequence — why the research lands the way it does — instead of on “then she talks about.”

Summary creep, and how to catch it

Summary creep is the slow drift from analyzing a text to retelling it, and it is the single most common failure in this assignment. It happens because retelling is easy and analysis is not, and because the sentences feel productive while you are writing them. Three diagnostics catch it reliably.

The subject test. Highlight the grammatical subject of every sentence in your draft. Subjects like “nurse staffing,” “the shortage,” “patients” are about the world. Subjects like “the third paragraph,” “the concession,” “Alvarez” are about the text. If more than roughly a third of your subjects are world-subjects, you have written an essay about staffing, not about an essay about staffing. The deletion test. Cut every quotation from a paragraph. If the paragraph still makes complete sense, you were not analyzing the quotation — you were decorating around it. The “so what” chain. After each sentence of description, force yourself to write “which matters because…” If you cannot finish it twice in a row, the passage is filler.

Words worth knowing