Skip to content

Rhetorical Skill Development

Principle · All principles
Evidence · 18 claims (14 for, 4 mixed) · 23 studies (7 causal, 6 quant-synthesis, 5 theoretical, 3 review, 1 associational, 1 qualitative), q1–q4 · 5 of 23 report an effect size · 13 claims rest on one study

Conditional relationship

This page owns language taught as a set of choices with consequences for an audience: in writing, speech, debate and everyday talk, deciding what to claim, what evidence to give, how to answer the other side, and which words, register and directness fit the occasion. For a learner who has ideas but presents them the same way to every reader (a novice or intermediate writer or speaker in the genre at hand: drafts with no reader in view, no answer to an objection, style handled as rules to obey), teaching rhetorical moves as decisions made for a stated audience, purpose and situation, with explicit criteria for evidence and for answering the other side, practice across contrasting situations, process goals before outcome goals, and learners saying why a move works, is expected to produce better arguments and texts or talk better fitted to their audience (skill-performance, near-transfer to a new topic or audience) than instruction in a fixed set of forms or prescriptive style rules. The target is a complex-skill with high element interactivity: evidence, stance, structure, wording and delivery have to be judged together.

No claim in the wiki tests this relationship as stated: none compares choice-centred rhetorical instruction with instruction in fixed tools or style rules. The claims below test neighbouring parts of it: practising argument and its transfer to essays on new topics, explicit strategy instruction for writing, writing for a real audience, structured peer feedback, self-explanation prompts and process goals. They come from different learners and settings (adolescents in multiyear dialogue programmes, grades 4–10, second graders and seventh graders, undergraduates, mathematics and science learners), and each is carried here only as far as the row that uses it says. Three thin pages were folded into this one in October 2026, and their recommendations are steps below: teach strategic choices rather than fixed tools (Cyphert 1996, a theoretical paper; no claim), learn when a direct and when an indirect message fits the situation (Duke & Nowicki 1979; one claim, describing variation, not testing teaching), and teach style as consequential choice rather than rules (Pace 2001; two claims, both a single author's argument and one class narrative).

How the converted siblings sit beside this: Authentic Audiences and Purposes owns the relationship between who the work is for and its quality; this page uses a real audience as one condition and does not repeat that model. Process-Based Writing owns planning, drafting, response and revision as a process; the drafting and revising steps here assume it. Epistemic Cognition owns judging claims by the standards of a kind of evidence, which is the substance a persuasive text must not lose. Self-Monitoring owns checks of one's own work against a criterion. The Debate, Structured Academic Controversy and Structured Peer Review patterns carry reusable policies for arguing a contested question and for reviewing drafts; use them for those segments.

Default design, while the relationship is untested

The earlier page's guidance and the three folded pages' recommendations, kept as a concrete default a designer can act on and revise. Each step is a design proposal; its evidence status is stated beside it.

  1. Start from a real rhetorical situation, not a free topic. Give each task an occasion, a named audience and a purpose (a letter to the department about a timetable change, a two-minute case to a local council, a reply to a skeptical colleague), so that what to say is decided by who will hear it. Untested as a rule; from the folded Cyphert page, which rejects topic selection apart from occasion, audience and speaker. Where the audience can be real, make it real: Authentic Audiences Improve Student Work [+M] (seventh graders and second graders wrote higher-rated texts for an outside audience than for the teacher; product quality, immediate).
  2. Choose issues that matter where the learners are, and model the first one. Use questions from the learners' own institution, workplace or community (the folded Cyphert page: students as participants in institutional decision-making). Learners do not at first know how to take on a difficult institutional issue, so the teacher frames and argues the first one aloud rather than setting a trivial practice exercise. Untested proposal, from a theoretical paper.
  3. Make the criteria explicit before anyone drafts. State what the work will be judged on: a clear claim, evidence that bears on it, the reason linking them, an answer to the strongest opposing view, an organisation the reader can follow, and wording fitted to this reader. Score evidence and reasoning separately from delivery and polish, so a smooth text without support does not pass. The set of criteria is an untested proposal; that practising argument improves argument quality is supported by Argumentation Improves Reasoning [+M] (online science argumentation in 46 studies; a multiyear dialogue programme whose adolescents wrote better essays on new topics than a comparison group given essay practice).
  4. Teach one move at a time, as a process goal. Model a move (anticipate the skeptical reader's question and answer it; open with what this audience already accepts), have learners practise it, and set the goal as the move ("include and answer one objection") rather than the effect ("convince the reader"). For novice learners keep every goal a process goal for the whole course: when a move is used without prompting, the next goal is the next move (or the same move for a harder reader), not "persuade the audience"; outcome goals are for learners who already adapt to readers (see the advanced-learner row). Process goals lead to better skill acquisition for novices than outcome goals. [+M] (high-school students learning writing revision; the study also suggests a shift to outcome goals once a skill is mastered, which a short course for novices rarely reaches). Explicit, modelled strategy teaching of this kind for planning and revising: Strategy Instruction Improves Writing Quality [+M] (one meta-analysis, grades 4–10; every study taught for at least three days, with modelling and progression toward independent use).
  5. Ask why a move works, not only for an imitation of it. After each modelled or peer example, the learner says or writes in a sentence what the move does for this reader ("the concession makes the council more likely to hear the request because…"). Carried from Prompting learners to self-explain improves understanding and problem solving on immediate tests by a moderate average amount, with little evidence yet on delayed or classroom gains. [+S] (mostly mathematics and science; untested for rhetoric, and the claim's own title limits it to immediate tests).
  6. Analyse style as choice, using a real text from the learners' setting. Take a text that matters locally (in the folded Pace page, a university president's letter after campus protests): learners examine its word choice and sentence structure, imitate one sentence, translate a passage into their own vernacular, and say what each choice does to its readers and for its writer. Teach a convention (formal register, the passive in a lab report) as a choice with an audience, not as a rule. Teaching style can be reclaimed as an effective tool for socially responsible rhetorical communication [+W] and In one first-year composition class, stylistic analysis of a president's letter served critical pedagogy goals [+W] (one author's argument and his account of one class; no outcome measured).
  7. Vary the situation and ask what fits. Give the same message for two situations that differ in how much is at stake and how close the relationship is (refusing a request from a close friend and from a manager; a complaint to a neighbour and to a landlord), and have learners decide how direct, how warm and how formal each version should be, and whether what they say should match what they feel. The folded Duke & Nowicki page: effective communicators can be both congruent and incongruent and know when each fits. Congruence of communication varies with situational importance and relationship closeness [~M] (60 female college students on a questionnaire answered more congruently when the situation and the relationship were important; it describes variation, it does not test teaching it).
  8. Rehearse at low stakes before the public or graded version. Each draft or run-through goes to one partner who comments against the criteria in step 3 (two comments: one strength, one change), and the writer or speaker revises before the final version. Grouping rule: work in pairs; with an even number of learners every learner has a partner and no group is uneven; with an odd number, form one group of three in which each member reviews one draft in a rotation (A→B→C→A), so every learner still gives and receives exactly one review; change partners between tasks so each learner meets several readers over the course. Peer Feedback Improves Writing [+M] (one meta-analysis of 24 higher-education studies; structured peer feedback against no feedback and against self-assessment; the table on that page gathers 4 distinct studies). The low-stakes ordering is an untested proposal from the earlier page.
  9. Have learners check their own draft against the criteria while revising. Three questions before submitting: does this reader know why it matters to them; is every claim backed by evidence they would accept; is the strongest objection answered? Self-regulated learning models treat self-monitoring as a core process through which learners notice gaps and adjust strategy [~M] (a review and a theoretical synthesis; it does not test self-checking in rhetoric).
  10. Ask learners to notice their choices outside the course. A short weekly log entry: one situation outside class where they chose words for an audience, what they chose and why. Some students expanded and generalized class tools and rhetorical concepts into their own activities beyond the course [+W] (two second-language writing students in a case study; one carried audience awareness into arguments with friends and relatives). The log is an untested proposal.
  11. Plan the arc across the sessions the brief gives, as fractions of the course. Whatever the number of sessions, divide them into three phases of about a quarter, a half and a quarter (with 12 sessions: 3, 6, 3; with 8: 2, 4, 2; with 4: 1, 2, 1; round so the middle phase is longest). All figures are untested proposals except the floor of three days of instruction that every study in the strategy-instruction meta-analysis met; below three sessions use the one-session row. Within a session, give about a third of the time to modelling or analysis and two thirds to learners drafting, responding and revising (in a 75-minute session: 25 and 50).
    • First quarter (modelling): the teacher brings two or three model texts or recordings from the learners' setting; analysing the choices in them (step 6), one move modelled aloud (step 4), practice on that move with the model visible. Process goal: the first move.
    • Middle half (practice across situations): one new move every one or two sessions, each set as a process goal; learners write or speak a message for contrasting audiences (step 7), get one partner's response against the criteria (step 8), and revise. The source of examples shifts from the teacher's texts to texts the learners find, and the teacher's prompts to the learner's own three-question check (step 9).
    • Last quarter (whole product): one product or performance for a real audience (step 1), drafted, rehearsed with a partner, self-checked and delivered. For novices the goals stay process goals ("use the three moves you have practised for this reader"); only learners already adapting to readers take an outcome goal (step 4).
    • Moving on: a learner moves to a new move, a harder audience or a less familiar genre once two consecutive tasks meet the evidence and opposing-view criteria without prompting and the learner can explain, for one choice, what it does for that reader. A learner whose drafts still lack evidence after two rounds goes back to a model text and one move, not on to style.

When a condition is not met: if no real audience is available, use a named, described reader who will actually read the piece (another class, a visiting practitioner) rather than "imagine the mayor"; if there is little time, cut scope rather than pace: one product, revised once, before a second product is attempted; if there is no time for revision, teach fewer moves and keep the rehearsal, since a performance with no chance to act on feedback tests rather than teaches.

Fitting the design to a situation

The three facts that most change the decision:

  • What the learner does now, on this kind of task. A learner who writes with no reader in view needs one move modelled and a process goal; one who already adapts to readers needs harder audiences and outcome goals, and scripted routines may get in the way. If unstated, ask: given a short task for a stated reader, does the learner change anything for that reader, and can they answer one objection?
  • Who the audience is, and whether it is real. A real reader who will respond changes what learners write; a token one may not. If unstated, ask: who will actually read or hear the final version, and will they reply or act on it?
  • The kind of goal. Argument quality in a subject, persuasive or professional writing, public speaking, and everyday interpersonal communication call for different tasks, criteria and situations. If unstated, ask: what must the learner produce, for whom, judged on what, and when?
If the brief says… Then change… Basis
Learners are novice or struggling writers Teach one move at a time with a modelled demonstration and a process goal, and keep goals as process goals to the end of the course (the next goal is the next move, not the audience's verdict); run strategy teaching over at least three sessions; keep the model text visible during practice Strategy Instruction Improves Writing Quality [+M] (grades 4–10; struggling writers gained more); Process goals lead to better skill acquisition for novices than outcome goals. [+M] (high-school students)
Learners already adapt to readers (advanced, experienced professionals) Drop scripted move-by-move routines; set outcome goals (the reader's response); give harder audiences, hostile questions and critique of published texts Process goals lead to better skill acquisition for novices than outcome goals. [~M] (suggests shifting to outcome goals after mastery; does not test advanced learners)
Learners are child (primary age) Write or speak for an outside reader they can picture (a pen-pal class, a local business); keep tasks short and oral first; the teacher models one choice at a time Authentic Audiences Improve Student Work [+M] (second graders wrote higher-quality texts for an external audience; within-subjects, abstract only). The oral-first step is untested
Learners are adult in work settings Use their own documents and occasions (a proposal to a manager, a client email, a briefing); take the workplace's own categories of document and reader; practise the same message for two levels of seniority untested proposal
Goal is interpersonal communication rather than argument Use paired situations differing in stakes and closeness (step 7); learners decide how direct, warm and formal each reply is, and when saying exactly what they feel fits Congruence of communication varies with situational importance and relationship closeness [~M] (describes variation among college women; does not test teaching)
Goal is argument quality in a subject (science, history) Make the subject's own standards of evidence the criteria; add structured peer dialogue on a contested question before individual writing Argumentation Improves Reasoning [+M] (online science argumentation; dialogue programme transferred to essays)
Learners join remotely, some online and some in the room, or on phones only Run argument and peer response as written exchanges in one shared thread, so remote and in-room learners take the same turns; keep each contribution to a few sentences on a phone Argumentation Improves Reasoning [+M] (studies of online and electronically conducted argument; the mixed-mode arrangement is untested)
One teacher with a large group Peer response in pairs against the written criteria (one group of three in rotation only if the number is odd); the teacher reads a sample of drafts and gives whole-class feedback on the commonest gap (usually the missing answer to an objection) Peer Feedback Improves Writing [+M] (higher education; carried to school classes untested)
Facilitator time is scarce (volunteers, one tutor checking in briefly) The tutor checks one thing per draft: is the strongest objection answered? Everything else goes to the learner's three-question self-check and to peers untested proposal
One session only, including one short online session Cut scope to one product revised once: one situation and reader, one move, one short script or paragraph. For a 30-minute session: 5 minutes on the reader and the criteria, 5 on one model choice, 8 drafting, 4 writing a one-sentence self-explanation of why the move suits this reader (a partner's comment instead, if there is a partner), 8 revising the same draft = 30. Do not draft or perform a second version for another audience; record what the revision shows and do not claim transfer untested; the strategy-instruction studies all ran at least three days (Strategy Instruction Improves Writing Quality [~M])
A fixed short time budget per session (for example 45 or 75 minutes) Keep one product per session and one revision of it; split the time about one third modelling, two thirds drafting, responding and revising, and write minutes that add up (75: 10 reader and criteria, 15 model choice, 20 drafting, 10 partner response, 5 self-explanation of one choice, 15 revising). Drop style analysis before dropping the revision untested proposal
The brief gives a number of learners Work in pairs: an even count makes all pairs; an odd count makes pairs plus one group of three reviewing in rotation; never leave a learner without a reviewer or set triads where pairs fit untested proposal
Learners have low confidence, low literacy or are writing in a second language Rehearse orally before writing; let learners plan in their first language; first rehearsal with one trusted peer, not the group; gloss the criteria in plain words untested; Some students expanded and generalized class tools and rhetorical concepts into their own activities beyond the course [+W] shows two second-language learners taking audience awareness into daily life, not a test of these supports
The work is a product or performance for a real audience Keep it whole: draft and rehearse the full speech or document, not separated exercises; rehearse with peers first; judge substance and delivery separately Authentic Audiences Improve Student Work [+M] (immediate product quality); the whole-task rule is untested
Stakes are high (graded public speech, competition, assessed portfolio) Put at least one low-stakes rehearsal with peer response before the graded version; grade evidence and reasoning separately from polish untested; from the earlier page ("performance pressure can narrow experimentation")
The brief names its own rubric, criteria or rhetorical categories Use them as given; add only an evidence check and an answer-to-objection check if the rubric lacks them, and teach the named categories as choices for a reader untested proposal
The topic is personal or sensitive (a learner's own stance, an institutional dispute, their community) The learner chooses the issue and who sees the work; positions are not required in public; analyse a public text rather than a classmate's views untested proposal; the folded Cyphert and Pace pages use institutional issues, without addressing privacy
A convention is mandated (exam essay form, house style) Teach the convention as one choice with its own audience (the examiner, the journal), alongside a contrasting version for a different reader Teaching style can be reclaimed as an effective tool for socially responsible rhetorical communication [+W] (theoretical argument)

Eleven of the eighteen rows cite a claim, but none tests rhetorical-choice teaching itself, and two (one session; low confidence) cite one only for a limit or an illustration; the rest carry a neighbouring finding (writing strategy instruction, audience effects on writing quality, online argumentation, peer feedback in higher education, a questionnaire on congruence, one author's argument about style) to this setting, and the remaining seven are untested proposals.

Observation, state and explanation

What can be observed is a draft or a performance for a stated audience under stated conditions: what the learner claimed, what evidence they gave, whether they answered an objection, what they changed between versions and after which comment, and what they say a choice does for the reader. "Has audience awareness", "understands style as choice" and "can argue" are inferred states. A text that persuades one rater shows that it worked on that rater, not that the writer chose its moves for that reason.

Response Plausible explanations Observation that could change the interpretation
Polished, fluent text with little evidence Style taught without substance; the learner does not know what counts as evidence in this subject; the criteria rewarded fluency Score evidence and delivery separately; ask the learner what a skeptical reader would ask for, and whether they could find it.
The same text for two different audiences Does not see that the audiences differ; sees it but lacks the words to change register; thinks changing it would be insincere Ask what each reader already knows and wants; give one sentence rewritten for one reader and ask for the other.
Uses a taught move everywhere, including where it does not fit The move was learned as a rule, not a choice; process goal kept too long Give a situation where the move would backfire and ask whether to use it and why.
Answers objections in writing, not in live debate Time pressure in live talk; anxiety in front of peers; the objections in writing were prepared Compare a prepared and an unprepared objection, in writing and aloud, in a small group.
Strong in class, no change in their writing elsewhere Classroom tasks too unlike the learner's own occasions; no prompt to notice choices outside class Ask for one outside-class example a week and compare it with class work.

These rows are proposals, untested with learners as diagnostics. Asking a learner to explain a choice also teaches it, so record each probe, and keep the explanations that remain open rather than choosing the one whose name fits.

Evidence and qualifications

No claim in the wiki compares teaching rhetorical choices with teaching fixed forms or style rules, and none follows rhetorical skill to a delayed outcome. The claims below test neighbouring parts of the relationship.

  • Argumentation Improves Reasoning [+M]: a meta-analysis of 46 studies of online scientific argumentation (5,415 students) reports medium effects on argumentation skill (g = 0.603) and on content learning through argument (g = 0.596), larger in K-12 than in higher education (abstract only); a multiyear quasi-experiment found that young adolescents who argued about social issues in electronic peer dialogues wrote better arguments on new topics, with more attention to evidence, than comparison groups given essay practice and whole-class discussion. It supports practising argument against criteria; it does not test audience adaptation or style.
  • Strategy Instruction Improves Writing Quality [+M]: one meta-analysis, 20 effect sizes in grades 4–10, of explicit, modelled instruction in planning, revising and editing strategies, at least three days long: a large weighted mean effect on writing quality (0.82), larger with self-regulation components (1.14 against 0.62) and for struggling writers (1.02). The claim's discussion says effects are strongest in persuasive and expository genres; no study involved students above grade 10.
  • Authentic Audiences Improve Student Work [+M]: seventh graders (44, two classes) wrote higher-rated compositions to distant peers than to the teacher for a grade, order counterbalanced; second graders wrote better for an external audience; a narrative review proposes the mechanism. The table on that page gathers 8 distinct studies, 3 of them causal, none with an effect size coded. Product quality at the time only.
  • Peer Feedback Improves Writing [+M]: one meta-analysis of 24 higher-education studies: structured peer feedback against no feedback g = 0.91, against self-assessment g = 0.33, against teacher feedback an interval crossing zero; giving and receiving are pooled. The table on that page gathers 4 distinct studies.
  • Prompting learners to self-explain improves understanding and problem solving on immediate tests by a moderate average amount, with little evidence yet on delayed or classroom gains. [+S]: two meta-analyses (one reporting g = .55 across 69 effect sizes) and two small studies, mostly mathematics and science. It is the basis for asking why a move works; carried to rhetoric untested.
  • Process goals lead to better skill acquisition for novices than outcome goals. [+M]: high-school students learning writing revision did better with process goals than outcome goals, and shifting to outcome goals after mastery appeared to help; a theoretical chapter synthesises similar findings. No effect size is printed.

What limits the model.

Learners, tasks and outcomes differ across these claims, so do not rank them by effect label. Most outcomes are product quality or argument quality at the time of the study; none measures retention or use of rhetorical choices months later.

The evidence by question

Further evidence, not yet read against this model

Evidence the page cites or touches that is not part of the model above.

Objective and learner-valued goal

The designer's objective is usually a stated product or performance for a stated audience (an argument essay, a speech, a professional document) judged on stated criteria, at a stated horizon. Elicit separately what the learner wants from it: to pass the speech assessment, to be taken seriously at work, to win arguments with family, to have a say in a decision that affects them, or to avoid speaking in front of others. Agreement and divergence are both observations. A learner who values not being embarrassed may avoid a real audience the design depends on; one who wants to change a decision may care about persuading a particular reader more than about the rubric.

For example, a first-year student's objective in a composition course is an argument essay that meets the course rubric, while the student wants to write a complaint to housing services that gets a response. Teaching the moves on the course essay serves the objective; one draft of the actual complaint, read against the same criteria and sent, tests the learner's goal. Record both, so that a strong essay is not read as the capability the student came for.

What would revise this model?

The expectation should weaken if comparisons of choice-centred teaching with fixed-form or rule-based teaching, with comparable learners and outcomes judged by readers blind to condition, show no advantage, or show that gains on taught tasks do not carry to a new audience or topic. If learners taught to vary register and directness by situation do no better in new situations than those who practised one form, step 7 should go. If style analysis of local texts adds nothing over strategy instruction alone, drop it to save time. If peer rehearsal at low stakes does not change the graded performance, the ordering in step 8 is a preference, not a design rule. Do not protect the model by calling every failure "not a real enough audience" or "too little practice" after the fact.

Argument quality on the taught task, adaptation to a new audience, use outside the course, and retention months later are separate claims. The present evidence establishes no number of sessions beyond the strategy-instruction studies' three-day floor, no moving-on threshold, and nothing about delayed outcomes.

Examples

Illustrative

Argument writing with audience revision — Learners revise a claim or essay after considering what a skeptical reader would question or need clarified.

Debate and rebuttal routines — Students practice using evidence strategically, anticipating counterargument, and adjusting language for persuasive effect.

Presentation critique with rhetorical criteria — Peers or instructors give feedback on clarity, organization, evidence use, and audience adaptation rather than on style alone.

Key Sources

  • Toulmin, S. (2003). The uses of argument (Updated ed.). Cambridge University Press. doi:10.1017/cbo9780511840005
  • Graff, G., & Birkenstein, C. (2021). They say / I say (5th ed.). W. W. Norton.
  • Cyphert, Dale. (1996). Taking the Helm in Critical Pedagogy: The Basic Speech Curriculum as an Operationalization of the Paradigm. https://eric.ed.gov/?id=ED403616
  • Duke, Marshall P. and Nowicki, Stephen. (1979). A Social Learning Theory Conceptualization of Interpersonal Relationships. https://eric.ed.gov/?id=ED183975
  • Pace, Tom. (2001). Composing and Community: The Teaching of Style as a Tool of Critical Pedagogy. https://eric.ed.gov/?id=ED451526
Page metadata
Field Value
type principle
id rhetorical-skill-development
aliases teach-rhetorical-choices-not-tools, flexible-congruence-context-appropriateness, style-as-consequential-choice-not-prescription, teach-persuasive-standards-not-just-persuasive-skill, teach-canonicity-not-canon, teach-popular-culture-not-speech-making, teach-civic-responsibility-not-proficiency
title Rhetorical Skill Development
description For a learner who presents ideas the same way to every reader, teaching rhetorical moves as choices made for a stated audience, purpose and situation, with explicit criteria for evidence and for answering the other side, practice across contrasting situations, process goals first and learners saying why a move works, is expected to produce better arguments and better-fitted texts or talk than teaching fixed forms or style rules; no claim here tests that comparison, though argument practice, writing strategy instruction, real audiences and peer feedback are each tested.
status review
canonical True
generated by claude/unspecified on 2026-10-14
trust unverified — no one has confirmed this page's content
sources 3 — cyphert-1996, duke-1979, pace-2001

Frontmatter is the page's machine-readable record — see the Schema & Guide.