Skip to content
Back to search results

Appraisal Analysis

Introduced by
J. R. Martin and P. R. R. White
Year
2005
Type
Qualitative analysis of evaluative and stance-bearing language
Family
Systemic functional linguistics / discourse semantics
Font size

What does it do?

Appraisal analysis is the systematic study of evaluative language — how speakers and writers express feelings, make judgements, value things, take a stance toward other voices, and turn the volume of their evaluations up or down.

Developed by James Martin and Peter White within the interpersonal metafunction of systemic functional linguistics, the appraisal framework codes evaluative meaning along three systems: ATTITUDE (the kinds of feeling expressed), ENGAGEMENT (how the text positions itself among alternative voices and viewpoints), and GRADUATION (how evaluations are intensified or softened, sharpened or blurred). The method makes the often-invisible work of evaluation explicit, showing how texts construe stance and build alignment with their readers.

The intuition behind the method

Texts are rarely neutral. Beneath the propositions, a steady current of evaluation tells us how the writer feels, who is praised or blamed, what is beautiful or shoddy, how certain a claim is, and whose voice is being admitted or shut out. Appraisal analysis is the instrument for tracing that current. It asks three questions: What is being valued, and as what kind of feeling — emotion, ethics, or aesthetics? How does the text engage with other possible positions — asserting flatly, hedging, attributing, conceding? And how strongly is each evaluation pitched? Reading a text through these systems reveals the rhetorical and ideological work that ordinary evaluative wording quietly performs.

History

The appraisal framework emerged in the 1990s from the 'Sydney School' of systemic functional linguistics, in work by James Martin, Peter White, and collaborators on the interpersonal semantics of evaluation, partly motivated by genre-based literacy projects analyzing how writers express and negotiate attitudes. Martin's 2000 chapter 'Beyond exchange: Appraisal systems in English', in Hunston and Thompson's Evaluation in Text, set out the systems of attitude, engagement, and graduation as a discourse-semantic extension of Halliday's interpersonal metafunction. The framework was consolidated in Martin and White's 2005 monograph The Language of Evaluation: Appraisal in English, which remains the standard reference. Since then appraisal has become one of the most widely applied tools for analyzing evaluative language across media studies, academic discourse, narrative, and computational stance and sentiment research.

When to use it

Use appraisal analysis when your question is about evaluation, stance, emotion, or alignment in texts — how a writer expresses feeling, praises or condemns people and things, manages certainty and the voices of others, and intensifies or downplays their assessments. It is well suited to studies of opinion and persuasive writing, media bias, academic and political stance, identity, and reader alignment. It is less appropriate when evaluation is not the focus (other discourse methods may serve better), when the corpus is too large for the labor-intensive manual coding the framework requires without sampling, or when borderline and invoked evaluations are so frequent that reliable coding becomes impractical.

Limitations and considerations

  • Coding is highly interpretive, especially for invoked (implicit) evaluation, so inter-coder reliability can be hard to achieve.
  • The detailed system network is complex and demands considerable training to apply consistently.
  • Manual coding does not scale easily to large corpora, forcing sampling that may limit generalization.
  • Category boundaries — for instance between affect and judgement, or between sub-types of engagement — can be genuinely indeterminate.

Source conditions do not replace independent scientific review or validation by an analysis engine.

Strengths

  • Provides a fine-grained, comprehensive taxonomy of evaluative meaning, going well beyond simple positive/negative sentiment.
  • Captures not just what is evaluated but how the text manages other voices and scales its commitments through engagement and graduation.
  • Grounded in systemic functional linguistics, it links evaluative wording to interpersonal meaning, genre, and ideological positioning.
  • Has proven adaptable across genres and languages, supporting comparative and discourse-analytic studies of stance and alignment.

Applications

  • Media and political discourse: analyzing bias, stance, and reader alignment in news reporting, editorials, and speeches.
  • Academic writing research: examining how writers express evaluation, hedge claims, and engage with other scholars' positions.
  • Identity and narrative studies: tracing how affect and judgement construct selves and characters in personal and literary texts.
  • Persuasion and marketing: studying how advertising and review texts use attitude and graduation to position audiences.
  • Computational stance and sentiment analysis: informing finer-grained models of evaluation that go beyond binary sentiment polarity.

How it works

How it works

1

The analyst first reads through the text to locate instances of evaluation — the words and phrases through which the author conveys positive or negative assessments, commitments, and positionings. Following Martin's account of appraisal as the discourse-semantic resource for negotiating attitudes and intersubjective stance, the analyst distinguishes inscribed evaluation, which is explicitly lexicalized (brilliant, corrupt, fortunately), from invoked or evoked evaluation, where evaluation is implied by ideational content or rhetorical strategy and read off by the disposed reader. The polarity (positive or negative) and the source of each evaluation (the author's own voice or an attributed voice) are noted.

2

Identified evaluations are coded for ATTITUDE, the system of feeling, which divides into three regions. Affect registers emotional responses (happiness, security, satisfaction, and their negatives — 'I was delighted', 'they feared'). Judgement evaluates human behaviour against social norms, split into social esteem (normality, capacity, tenacity — 'a skilled, determined leader') and social sanction (veracity and propriety — 'a dishonest, corrupt official'). Appreciation evaluates the worth of things, texts, and phenomena along reaction, composition, and valuation ('a moving, well-crafted, important film'). The analyst tags each instance with its subtype and polarity.

3

The analyst then codes ENGAGEMENT and GRADUATION. Engagement, drawing on Bakhtin's dialogism, captures how the text positions itself relative to other voices: monoglossic utterances present a proposition as bare assertion with no alternatives acknowledged, while heteroglossic utterances open or close dialogic space — entertaining a position (it may be, perhaps), attributing it to a source (she claims, reportedly), proclaiming it (indeed, the facts show), or disclaiming it through denial or counter-expectation (not, however, although). Graduation, which cuts across the whole system, codes the scaling of evaluation: force grades intensity and amount (slightly, very, a huge number), and focus sharpens or softens category membership (a true friend, kind of, sort of).

4

Finally the analyst synthesizes the coded patterns into an account of the text's evaluative stance and its strategy for aligning readers. By tracing prosodies of attitude across a text, the dominant engagement strategies, and where graduation peaks, the analyst shows how the writer construes an authorial persona, builds a community of shared values, opens or forecloses debate, and rhetorically positions the audience. These findings connect to genre, register, and ideology, making appraisal analysis a key method for studying stance, persuasion, identity, and bias in discourse.

Common pitfalls

  • Coding only inscribed (explicit) evaluation and missing the invoked evaluation that carries much of a text's attitudinal work.
  • Confusing affect, judgement, and appreciation — for example tagging an evaluation of a person's character as appreciation rather than judgement.
  • Treating engagement as mere modality and overlooking its dialogic dimension of opening and closing space for other voices.
  • Counting attitude tokens in isolation without tracing the evaluative prosody that builds across stretches of the text.

Frequently asked questions

How does appraisal analysis relate to systemic functional analysis?

Appraisal is a specialized extension of systemic functional analysis within the interpersonal metafunction. Where general SFL analysis treats interpersonal meaning mainly through mood and modality at the level of the clause, the appraisal framework develops a richer, discourse-semantic account of evaluation that runs across stretches of text — the systems of attitude, engagement, and graduation. So appraisal analysis is best understood as a focused tool for the evaluative dimension of meaning, sitting inside the broader SFL programme and often used alongside transitivity and mood analysis.

What is the difference between affect, judgement, and appreciation?

These are the three regions of the ATTITUDE system. Affect deals with emotional reactions — how people feel ('I was thrilled', 'they were anxious'). Judgement evaluates human behaviour and character against social norms, covering social esteem (admiration of capacity, tenacity, normality) and social sanction (approval or condemnation of veracity and propriety, as in 'honest' or 'corrupt'). Appreciation evaluates the value of things, products, performances, and phenomena, through reaction, composition, and valuation ('beautiful', 'well-structured', 'significant'). The practical test is what is being evaluated: a feeling, a person's conduct, or the worth of a thing.

What is engagement and why does it matter?

Engagement, drawing on Bakhtin's dialogism, describes how a text positions itself among other possible voices and viewpoints. A monoglossic statement presents a proposition as a bare fact with no alternatives recognized, whereas heteroglossic resources acknowledge other positions — entertaining them ('perhaps'), attributing them to a source ('she claims'), proclaiming agreement ('indeed'), or disclaiming them through denial or counter-expectation ('not', 'however'). Engagement matters because it reveals how a writer opens or closes space for dialogue, manages solidarity and disagreement, and rhetorically aligns or distances the reader — work that simple sentiment or modality analysis cannot capture.

Related methods

Sources

  1. Martin, J. R., & White, P. R. R. (2005). The Language of Evaluation: Appraisal in English. Palgrave Macmillan.

    ISBN: 9781403904096
    Copied.
  2. Martin, J. R. (2000). Beyond exchange: Appraisal systems in English. In S. Hunston & G. Thompson (Eds.), Evaluation in Text (pp. 142–175). Oxford University Press.

    ISBN: 9780198238546
    Copied.