On understanding computers and cognition

Podcast

---

Central Problem

How can a hermeneutic-phenomenological theory of language provide better foundations for computer technology design than the rationalistic tradition, and how can this theory account for the different interpretations that arise from the same text among readers with different backgrounds?

Main Thesis

Language does not convey information but evokes understanding through an interaction between utterance and the listener’s pre-understanding — a background of concerns, practices, and breakdowns generated within a tradition. This background-dependent theory of interpretation explains why the same book produced radically different readings among reviewers, and provides the foundation for designing computer systems as tools for human commitment and action rather than as representations of knowledge.

Historical Context

Writing in 1987 as a response to reviews of their influential Understanding Computers and Cognition (1986), Winograd and Flores defend their critique of the rationalistic tradition in AI and cognitive science. The mid-1980s marked a period of both AI optimism (expert systems boom) and growing disillusionment with symbolic AI’s foundational assumptions. Drawing on Continental philosophy (Heidegger, Gadamer), speech act theory (Searle, Austin), and ethnomethodology (Suchman), they challenge the prevailing representationalist paradigm and advocate for a “language/action perspective” on computer design, exemplified by their commercial product “The Coordinator.”

Philosophical Lineage

flowchart TD
    Heidegger --> |thrownness, readiness-to-hand| Background[Background of Pre-understanding]
    Gadamer --> |hermeneutics, tradition| Background
    Wittgenstein --> |language games, forms of life| CritiqueRep[Critique of Representationalism]
    Searle --> |speech acts, commitment| LangAction[Language/Action Perspective]
    Background --> WinogradFlores[Winograd & Flores]
    CritiqueRep --> WinogradFlores
    LangAction --> WinogradFlores
    Ethnomethodology --> |situated action| WinogradFlores
    WinogradFlores --> Coordinator[The Coordinator System]
    WinogradFlores --> PostCognitivistHCI[Post-Cognitivist HCI]
    Suchman --> |plans and situated actions| WinogradFlores

    class Heidegger,Gadamer,Wittgenstein,Searle,Suchman internal-link;

Key Thinkers

Thinker	Dates	Movement	Main Work	Core Concept
Heidegger	1889–1976	Phenomenology	Being and Time	Thrownness, readiness-to-hand, breakdown
Hans-Georg Gadamer	1900–2002	Hermeneutics	Truth and Method	Tradition, pre-understanding, fusion of horizons
Wittgenstein	1889–1951	Ordinary Language Philosophy	Philosophical Investigations	Language games, meaning as use
Searle	1932–2025	Speech Act Theory	Speech Acts	Illocutionary acts, commitment
Suchman	1951–	Ethnomethodology	Plans and Situated Actions	Situated action vs. planning
Kuhn	1922–1996	Philosophy of Science	The Structure of Scientific Revolutions	Paradigm shifts, serious listening

Key Concepts

Concept	Definition	Related to
Pre-understanding	The background of concerns, practices, and history that shapes interpretation before any explicit understanding occurs	Thrownness, tradition
Background	Not a “set of beliefs” but the lived context of practices and breakdowns that generates possibilities for interpretation	Pre-understanding, tradition
Tradition	Shared history of conversations that shapes language and thought; not a definable “school” but a ground on which we work	Background, culture
Breakdown	Disruption in transparent practice that reveals what was previously taken for granted	Readiness-to-hand, design
Listening	Active interpretation shaped by background; different listeners hear different meanings from the same utterance	Pre-understanding, openness
Commitment	Language act in which one allows others to anticipate future actions; the basis of coordination	Speech acts, The Coordinator
Readiness-to-hand	Heidegger’s term for transparent tool use; what good design achieves	Breakdown, thrownness
Rationalistic tradition	Western intellectual heritage assuming thought can be reduced to logical manipulation of explicit representations	Representationalism, AI

Authors Comparison

Theme	Winograd & Flores	Traditional AI	Ethnomethodology
View of language	Evokes understanding through background	Conveys information through symbols	Constitutes social reality
Role of representation	Emerges from breakdown, not foundation	Foundation of cognition	Insufficient for capturing practice
Design approach	Language/action perspective	Knowledge representation	Systematic study of practices
Background treatment	Constitutive, unarticulable	Reducible to explicit beliefs	Observable through methodology
Computer’s role	Tool for commitment and conversation	Intelligent agent	Artefact embedded in practice

Influences & Connections

• Draws from: Heidegger (thrownness, readiness-to-hand, breakdown), Hans-Georg Gadamer (hermeneutics, tradition), Wittgenstein (language games), Searle (speech acts), Suchman (situated action)
• Responds to: Rationalistic tradition in AI, cognitivism, expert systems optimism, representationalism
• Influences: Computer-Supported Cooperative Work (CSCW), third-wave HCI, post-cognitivist design, workflow systems
• Critique of: Symbolic AI, knowledge representation, naive technological optimism, detached methodology

Summary Formulas

1. Language ≠ Information Transfer: Language evokes understanding through interaction with pre-understanding, not transmission of content
2. Tradition Thesis: Thought and language are shaped by shared history that cannot be chosen, designed, or precisely defined
3. Background Irreducibility: Background cannot be articulated as “a set of beliefs, desires, and dispositions” — it is lived, not represented
4. Design as Commitment: Computer systems should be tools for making and tracking commitments, not simulations of intelligence
5. Serious Listening: Understanding requires looking for how “apparent absurdities” make sense, not judging logical arguments

Timeline

• 1958: Wittgenstein’s Philosophical Investigations published (posthumously, English edition)
• 1960: Hans-Georg Gadamer publishes Truth and Method
• 1969: Searle publishes Speech Acts
• 1972: Terry Winograd publishes Understanding Natural Language (early AI work)
• 1976: Joseph Weizenbaum publishes Computer Power and Human Reason
• 1986: Winograd & Flores publish Understanding Computers and Cognition
• 1986: Suchman publishes Plans and Situated Actions
• 1987: Winograd & Flores publish this response to reviewers
• 1986–1990s: The Coordinator deployed as commercial workflow system

Notable Quotes

“Language does not convey information. It evokes an understanding, or ‘listening,’ which is an interaction between what was said and the preunderstanding already present in the listener.” — Winograd & Flores

“We participate in a tradition and it changes through our participation. But we do not choose it or design it. It would be foolish to ignore the power of this particular tradition because it cannot be precisely defined.” — Winograd & Flores

“This book is anti-illusion, not anti-technology.” — William Clancey (quoted approvingly by Winograd & Flores)

---

Warning

This annotation was normalised using a large language model and may contain inaccuracies. These texts serve as preliminary study resources rather than exhaustive references.