程序代写代做代考 arm Excel flex chain interpreter G. E. M. Anscombe’s Intention is a classic of twentieth-century philosophy.

G. E. M. Anscombe’s Intention is a classic of twentieth-century philosophy.
The work has been enormously influential despite being a dense and largely
misunderstood text. It is a standard reference point for anyone engaging
with philosophy of action and philosophy of psychology.

In this Routledge Philosophy Guidebook, Rachael Wiseman:

• situates Intention in relation to Anscombe’s moral philosophy and
philosophy of mind

• considers the influence of Aquinas, Aristotle, Frege and Wittgenstein on
the method and content of Intention

• adopts a structure for assessing the text that shows how Anscombe
unifies the three aspects of the concept of intention

• considers the influence and implications of the piece whilst distinguish-
ing it from subsequent work in the philosophy of action.

Ideal for anyone wanting to understand and gain a perspective on Anscombe’s
seminal work, this guide is an essential introduction, useful in the study of
the philosophy of action, ethics, philosophy of psychology and related areas.

Rachael Wiseman is Addison Wheeler Research Fellow at Durham Uni versity,
UK.

Routledge Philosophy Guidebook to

Anscombe’s Intention

ROUTLEDGE PHILOSOPHY GUIDEBOOKS
Edited by Tim Crane and Jonathan Wolff

University of Cambridge and University College London

Plato and the Trial of Socrates Thomas C. Brickhouse and Nicholas D. Smith

Aristotle and the Metaphysics Vasilis Politis

Rousseau and the Social Contract Christopher Bertram

Plato and the Republic, second edition Nickolas Pappas

Husserl and the Cartesian Meditations A.D. Smith

Kierkegaard and Fear and Trembling John Lippitt

Descartes and the Meditations Gary Hatfield

Hegel and the Philosophy of Right Dudley Knowles

Hegel and the Phenomenology of Spirit Robert Stern

Berkeley and the Principles of Human Knowledge Robert Fogelin

Aristotle on Ethics Gerard Hughes

Hume on Religion David O’Connor

Leibniz and the Monadology Anthony Savile

The Later Heidegger George Pattison

Hegel on History Joseph McCarney

Hume on Morality James Baillie

Hume on Knowledge Harold Noonan

Kant and the Critique of Pure Reason Sebastian Gardner

Mill on Liberty Jonathan Riley

Mill on Utilitarianism Roger Crisp

Spinoza and the Ethics Genevieve Lloyd

Heidegger on Being and Time, second edition Stephen Mulhall

Locke on Government D.A. Lloyd Thomas

Locke on Human Understanding E.J. Lowe

Derrida on Deconstruction Barry Stocker

Kant on Judgement Robert Wicks

Nietzsche on Art Aaron Ridley

Rorty and the Mirror of Nature James Tartaglia

Hobbes and Leviathan Glen Newey

Wittgenstein and the Tractatus Michael Morris

Aristotle and the Politics Jean Roberts

Merleau-Ponty and Phenomenology of Perception Komarine Romdenh-Romluc

Frege on Sense and Reference Mark Textor

Kripke and Naming and Necessity Harold Noonan

Kant on Religion within the Boundaries of Mere Reason Lawrence Pasternack

Wittgenstein and On Certainty Andy Hamilton

Aristotle and the Poetics Angela Curran

Spinoza on Politics Daniel Frank and Jason Waller

Anscombe’s Intention Rachael Wiseman

Routledge Philosophy Guidebook to

Anscombe’s Intention

Rachael

Wiseman

First published 2016
by Routledge
2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN

and by Routledge
711 Third Avenue, New York, NY 10017

Routledge is an imprint of the Taylor & Francis Group, an informa business

© 2016 Rachael Wiseman

The right of Rachael Wiseman to be identified as the author of this work has been
asserted by her in accordance with sections 77 and 78 of the Copyright, Designs
and Patents Act 1988.

All rights reserved. No part of this book may be reprinted or reproduced or utilised
in any form or by any electronic, mechanical, or other means, now known or
hereafter invented, including photocopying and recording, or in any information
storage or retrieval system, without permission in writing from the publishers.

Trademark notice: Product or corporate names may be trademarks or registered
trademarks, and are used only for identification and explanation without intent
to infringe.

British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library

Library of Congress Cataloging in Publication Data
Wiseman, Rachael.
Routledge philosophy guidebook to Anscombe’s Intention / Rachael Wiseman.
pages cm.—(Routledge philosophy guidebooks)
Includes bibliographical references and index.
1. Anscombe, G. E. M. (Gertrude Elizabeth Margaret). Intention. 2. Intention
(Logic) I. Title.
BC199.I5A539 2016
160—dc23
2015035839

ISBN: 978-0-415-82186-5 (hbk)
ISBN: 978-0-415-82187-2 (pbk)
ISBN: 978-1-315-65118-7 (ebk)

Typeset in Garamond
by Book Now Ltd, London

Contents

Acknowledgements xi

Introduction 1

1 Elizabeth Anscombe and Intention 10
Anscombe’s times 11
Oxford philosophy 13
Anscombe’s faith 17
Intention’s influence 20
Notes 24
Suggested reading 24

2 Three tasks for Intention 26
The genesis of Anscombe’s Intention 28
Modern moral philosophy 31
Intended and foreseen consequences 37
Intention with which 39
An account of action 40
Notes 45
Suggested reading 46

CONTENTSviii

3 ‘The subject under three heads’ 47
The subject introduced under three heads 48
The ‘connective’ approach 49
‘We are in the dark about the character of the concept’ (§1) 56
What is this description for? 63
‘A tool for the philosophy of action’ 65
The structure of Intention 72
Notes 75
Suggested reading 76

4 Intentional action (§§5–19) 77
The question ‘Why?’ (§§5–18) 78
‘It was involuntary’ and reason vs cause (§§5–11) 81
Known without observation (§8) 83
Involuntary movements (§§7–11) 92
Intention, motive, cause, reason (§§10–14) 97
Voluntary vs intentional actions (§§17–19) 106
Notes 109
Suggested reading 111

5 Intention with which (§§20–40) 113
Intentional actions, further intentions in acting,
intentions for the future (§§20–21) 114
Doing X in order to do Y (§22) 120
Doing Y in doing X (§§23–28) 124
Non observational knowledge? (§§29–33) 132
Practical reasoning as ‘ordinary reasoning’ (§§33–34) 138
The idea of logical compulsion (§33) 141
Wanting (§§34–40) 145
Notes 147
Suggested reading 150

6 The character of the concept of intention (§§42–49) 151
Form of description of events (§§46–48) 152
Diagnosing the Cartesian impulse 158
Practical knowledge (§§28–32, §§45–48) 161
Notes 175
Suggested reading 176

CONTENTS ix

7 Expressions of intention for the future
(§§2–3 and §§50–52) 177
The place of expressions of intention for the future
in Intention 178
Species of prediction (§§2–3) 179
‘I am going to do it unless…’ 186
Intention and the future 189
Notes 192
Suggested reading 193

Bibliography 194
Index 201

ACknowledgements

Thank you to Addison Wheeler, Durham University, and to the
Wheeler family, for the extraodinary fellowship that has enabled
me to finish this book. Mr Wheeler recognised that researchers
at the start of their intellectual life need nothing more than free-
dom to pursue ideas, and so created a fellowship that requries
only that one make ‘efforts for increased knowledge of people and
their make-up so as to enable them to make better use of their life
here on earth’. It is hard to think of a contemporary philosopher
whose work better fits that description than Elizabeth Anscombe.
Thank you also to Tom Stoneham and the University of York,
who allowed me to spend a term at Chicago University during my
teaching fellowship, and to the AHRC for funding the postgradu-
ate work with which my interest in Anscombe began. Thanks too
to Adam Johnson at Routledge for balancing perfectly patience
and pressure. Detailed comments from two reviewers have, I
hope, helped me to improve the manuscript in the final stages.

Many conversations have helped me with this book. I have
benefited more than I can say from discussions with members
of the Praktisches Denken und Gutes Handeln network, chaired
by Mattais Haase, Thomas Hoffman and Tim Henning. Most
of what is good in contemporary work on Anscombe has come

ACKNOWLEDGEMENTSxii

from members of this network. Adrian Haddock, Jen Hornsby,
Mattais Haase and David Hunter have been especially generous
with their time and patient with my Wittgensteinianism, and
have given me leg-ups when I needed them. At Durham, Andy
Hamilton has chivvied me along, discussed Wittgenstein with me
at length, and reminded me that philosophy is both fun and seri-
ous. Clare MacCumhaill and Luna Dolezal have been wonderful
interlocutors and good friends. I began this book at York. Bob
Clark was a mentor and friend from the start, and helped me to
understand Wittgenstein. Amber Carpenter showed me how to
reconnect analytic philosophy and real life; without her guidance
and friendship I would certainly not have been awarded my cur-
rent fellowship, and my philosophical horizons would be much
restricted. Marie McGinn is the best philosopher I know and has
been the greatest influence on my work. This book would have
been better if written under her guidance, but without her help
and support in the past, it would have been impossible. The good
parts are down to her.

On a personal note, I would like to thank Bill and Yvonne
who rescued me from a Craigslist catastrophe in Chicago – here’s
the book I told you I was writing! With the Masters at Gateshead
Synchronised Swimming Club – Caroline, Cait, Debbie, Ilana,
Rachel, Sally, and our intrepid coach Wayne – I have learnt all
about what it is to lack bodily awareness. My family have put up
with being snapped at every time they asked how the book was
going; thank you and sorry. Charlotte has kept me company in
the library and the pub. Ben, Kadie and little Wynn are wonder-
ful people and I am extremely grateful to have them in my life.
Most of all, Joseph, who alone knows how hard it was to write this
book. Thanks for waiting.

INTRODUCTION

Writing in the Times Literary Supplement in 2010, Jonathan Dancy
remarked that though Intention is ‘an acknowledged classic’ and
‘we need a study guide to help us through it’, the fact that no one
has written one is ‘not so amazing’:

Anscombe’s thought is so idiosyncratic and so personal, she does
things in quite her own way, she doesn’t give the reader any help at
all, she doesn’t appear to have an overall position and she doesn’t try
to put things in any kind of order – all these things make the task of
characterising her philosophical output daunting.

(Dancy, 2010, p. 28)

While Dancy is quite right that Anscombe gives her reader very
little help, and that the task ahead is daunting, he also makes a
mistake that is common in readers of Anscombe: that of thinking
that her work is neither unified nor systematic.

In fact, over her long career, Anscombe methodically describes,
with great care and attention, in different but overlapping contexts,

INTRODUCTION2

the ‘look of human life containing rules’ (Diamond, 1991, p. 5).
Because her topic is the ‘natural history of man’ (Anscombe,
undated b, p. 224), concepts and questions recur in different
contexts, approached from different directions. It is Anscombe’s
ability to illuminate, piece by piece, the parts of human life – for
her, also a Christian life – while at the same time revealing just
how much of what is left remains in darkness, that makes what she
says at once brilliant and absurdly difficult.

A distinction taken for granted in a paper on ethics becomes
the object of enquiry in one on sensations. A concept that seemed
unproblematic in a discussion of causation becomes deeply
puzzling in a paper on the memory. An aside about a class of invol­
untary acts becomes central to a discussion of sin. When one is
trying to get to grips with Anscombe’s thought, often the question
one can’t answer in text A is addressed somewhere deep in text
B, if only one knew to look there; but if one did one would find
Anscombe there showing that something that seemed simple and
clear in text C is, when viewed from a different angle, dreadfully
puzzling. As over 170 of Anscombe’s articles have been published,
and reading any one of them with proper understanding requires
many hours (weeks? months?) of work, attaining an overview of
her thinking – itself an overview of the order that is distinctive of
human life – is incredibly challenging.1

Anscombe’s work is about describing an order, but it is not
the sort of order that can be ordered: parcelled up, summarised,
simplified for a reader. Indeed, as we will see, Anscombe’s philo­
sophical – and indeed personal – aim is to show that what appears
simple or easy is in fact complex and difficult. Her work exposes
banality and shallowness when it masquerades as sophisticated or
deep thought, and aims to reveal that which is often hidden by
cliché and posturing. The directions Anscombe’s enquiries take
are dictated not by the structure of deductive argument, proposi­
tion following proposition, nor by the routes that have been taken
by philosophers who have gone before her, but the order that is in
the concept that is her topic.

These opening remarks hint at two of the central interpretative
claims in this Guidebook. The first is that Intention is a polished

INTRODUCTION 3

work with its own internal structure, and is complete by its own
lights. It is not, as one early reviewer complained, a ‘first draft’
(Heath, 1960, p. 282). For a number of years, it has been common
to treat Intention as a preliminary sketch of the view developed with
much greater sophistication by Donald Davidson; this Guidebook
views this as a serious error.2 A core task will be to reveal the order
that is there in Intention, an order that is dictated by the con­
cept that is its topic, and to show that though Anscombe carefully
delimits the scope of her study – for example, mental causation
and knowledge of limb position are said not to come under her
topic (see §11, p. 18; §28, p. 50)3 – her results are unambiguous
and do not await further development or technical embellishment.

The second interpretative claim is that Anscombe’s treatment
of intention is non-psychological. By this I mean: Intention does
not seek to describe a state of mind or to give an account of the
psychological processes or mechanisms associated with intention.
Instead her book describes the logic – or grammar – of a psycho­
logical concept. Intention’s task, as Frege said of his study of
logic, is ‘an investigation of the mind; of the mind, not of minds’
(Frege, 1956, p. 308).4

Frege studied the mind by formally representing the laws of
thought. He insisted that the laws of thought are not like laws of
nature; they are not ‘psychological laws in accordance with which
[mental process] takes place’ (Frege, 1956, p. 289). So, for Frege,
when one is studying the laws of thought one is not interested
in how humans in fact think and reason – it is not relevant, for
example, that in general humans will infer p from p&q nor that
attempting to entertain a contradiction causes psychological dis­
comfort in all reasonable people. Nor is he concerned with the
psychological processes – be they in the mind or in the brain,
conscious or sub­conscious, personal or sub­personal – which
attend thinking and reasoning and which one might discover by
introspection or by empirical study. It is of no interest to a study
of the laws of thought that one does or does not feel a transition of
thought in moving from antecedent to consequent to conclusion,
nor that brain process XYZ is always found to attend an act of
judgement. Rather, to describe the laws of thought is to describe a
formal order in which anything, if it is a thought, will participate.

INTRODUCTION4

This is what Frege means when he says he is undertaking a study
of the mind not of minds.

The Fregean idea of a law of thought takes on a distinctive form
in Anscombe’s work, as it did in the later Wittgenstein’s. This is
because where Frege views language – the medium of thought –
as an abstract symbolism, Anscombe and the Wittgenstein of the
Philosophical Investigations have a conception of language as a
communicative tool, whose logic is shaped by and shapes the lives
and actions of people. This organic conception of language means
that the task of studying the mind not minds requires one to attend
to more than laws of logic; one must take as one’s object human
life shaped by norms.5

When Anscombe examines and describes a human life shaped
by the concept of intention, it can appear that she is talking about
individual minds, but she is not. Her book, like Wittgenstein’s
Philosophical Investigations, contains numerous vignettes: a man
is doing the shopping for his wife (§32), another is taking part
in a conspiracy to murder some bad men (§23), some Nazis are
caught in a trap and set about planning a massacre (§38), a stu­
dent plans to fail an exam (§2). Anscombe invites us to imagine
asking the agent ‘Why are you doing that?’ and then to consider
what the implications of different answers to that question might
be. Which answers would be intelligible? Which would ring false?
Which would elicit further enquiries? This can make it appear as
if she wants to show us something of the psychology of the pro­
tagonists, about what is going on in their minds. But it is rather
the case that this question allows her to display features of the
concept of intention. As she says in Intention: ‘if [this] account
were supposed to describe actual mental processes, it would in
general be absurd. The interest of the account is that it describes
an order which is there whenever actions are done with inten­
tions’ (§42, p. 80). This contrast – between a description of actual
mental processes, and a description of an order that belongs to the
concept of intention – is at the heart of Intention, and is one that
any adequate interpretation will need to make central.

More of this Guidebook than one might expect – or perhaps wish –
will turn on questions about how to read Anscombe’s monograph,

INTRODUCTION 5

about the book’s method and about the status of the philosophical
insights contained therein. Writers have not raised these questions
in relation to Anscombe’s work in the way that they have in rela­
tion to Wittgenstein’s. This is a major oversight and one which
this Guidebook can only begin to rectify.

It may be that because Anscombe – unlike Wittgenstein – wrote
papers that look like regular philosophical discourse, contain rec­
ognisable arguments, and engage with particular philosophical
theories, that commentators have assumed that there is nothing
distinctive about her method of philosophising. It is sometimes
suggested that Intention may exemplify a ‘grammatical investiga­
tion’ but little attention has been paid to what that means, and
to what it means for the account of intention that the book con­
tains.6 It is true that Aristotle and Aquinas are clearly major figures
in Intention, and that the account of action that the book contains
is rightly described as Aristotelian or as Thomist. But it is never­
theless the case that Anscombe brings their thought to us in a way
that is distinctive of someone writing after the so­called linguistic
turn, and under the influence of Wittgenstein. As such, one can­
not hope to understand Anscombe’s Intention – much less her
work as a whole – if one does not first describe what kind of philo­
sophical investigation she is undertaking, the method she is using,
and why that method is (given the nature of the investigation)
the right one.

In a sense, once one understands the character of the investi­
gation – the kind of philosophy that Anscombe is doing in
Intention – the book itself becomes manageable. It is not technical,
the arguments are not overly complex, there is no philosophical
terminology, no background theoretical knowledge is required,
no literature survey necessary. This is not to say that it is at all
straightforward; Anscombe is not a philosopher who wishes to
‘spare other people the trouble of thinking’ (Wittgenstein, 1958,
p. viii). The difficult thing about Intention is not the arguments
or the concepts, but the difficulty of paying close and unwaver­
ing attention to what Anscombe actually says rather than what
one suspects that she must mean. The description she puts before
us is ‘enormously complicated’ (§43, p. 80) because what we
do with and in language is enormously complicated. It is also a

INTRODUCTION6

description that we feel cannot be correct. Wittgenstein suggested
that philosophers should greet each other: ‘Take your time!’
(Wittgenstein, 1980, p. 80). This is the best advice I can offer to
a reader of Intention.

A couple of remarks about the content and structure of this
Guidebook.

Intention, I suggest, is a book that requires a high­level inter­
pretation; before looking at the detail of the book we need to
pose a prior question about what Anscombe is up to. Compare
Wittgenstein’s Tractatus Logic-Philosophicus. Before beginning one
must decide how to read the book. Is it a critique of metaphysics
(the ‘austere reading’) or a piece of metaphysics (the ‘traditional’
reading)? Are the propositions intended to be metaphysical, gram-
matical, or nonsensical (or some combination of the three)? Is the
tone literal or ironic? Answering these questions requires careful
attention to the text, but also to the background to the text, to
things not included in the book itself: to what or to whom was
Wittgenstein responding? Who did he have in mind as his reader?
What problem was he addressing? What results did he expect?
For this reason philosophers seeking to understand the Tractatus
take into account Wittgenstein’s preface to his book, his famous
letter to the editor, his response to Russell’s introduction, his later
renunciation of its method, and so on. High­level interpreta­
tive decisions frame the reading and provide guidance at points
of local interpretative dispute. In many cases it will simply be
impossible to resolve an interpretative difficulty or dispute with­
out looking outside the text to a point about method, aim, or
context. The same is true of Intention, but where whole journals
have been filled with debates about the method and target of the
Tractatus, very little reflection has been given to the method and
target of Intention.

For this reason, Chapters 2 and 3 address the genesis and method
of Intention respectively. Chapter 2 sets out three specific ways in
which Intention responds to Anscombe’s injunction in ‘Modern
Moral Philosophy’ that moral philosophy ‘should be laid aside …
until we have an adequate philosophy of psychology’ (Anscombe,
1958b, p. 169), and places this injunction in the context of

INTRODUCTION 7

Anscombe’s Catholicism. Chapter 3 describes the method of
Intention, and sets out its overall structure and argument.7

Chapters 4, 5, 6 and 7 follow the text of Intention in a fairly
linear manner, using the ‘three heads’ under which Anscombe
introduces the topic to divide up these core chapters, with inten-
tion with which spanning two chapters. These should all be read
alongside the relevant passages in Intention, indicated at each
chapter’s head. In Intention the first head – ‘expression of inten­
tion for the future’ – is split between §§2–4 and §§50–52. For
reasons to do with exposition, I have chosen to discuss this head
after intentional action and intention with which.

I should highlight a danger in choosing to set things out as I
have. As will become clear, the key to understanding Intention
is to recognise that its topic is a single formal structure in which
descriptions of events can participate, either by occurring in
accounts of what will happen, of what is happening, or of what
has happened. Though, as Anscombe says, ‘there is nothing
wrong with taking the topic piecemeal’ (§1) she soon reveals
that the concept of intention utterly resists a piecemeal treat­
ment (see, especially §20 and §23). Commentators have tried
to take Anscombe at her word, but the prevalence of the view
in philosophy of action that the piecemeal approach contains
pieces which are themselves topics that can each be investigated
in isolation from the others, makes the strategy more perilous,
so far as understanding Intention goes, than Anscombe herself
perhaps realised.

My concern in this Guidebook is with exegesis and not with
critique. Although Anscombe liked an argument, the task of pre­
senting Intention in a manner such that it is clear what one would
have to be arguing against, is sufficiently challenging to occupy a
Guidebook. If I am right about the book’s method and aims, then
the question of how one should respond critically to her account is
by no means straightforward to answer. It looks like responses that
reject her view as false, or her arguments as invalid or unsound,
will not properly engage with what she says. The sort of objection
that will be fitting will either have to go much deeper, and take
issue with Anscombe’s metaphilosophy, or it will need to show

INTRODUCTION8

that Anscombe has misdescribed the linguistic practice that is her
topic. In an important sense, Intention aims to say only what ‘isn’t
a philosophical thesis at all, and which no one denies’ (Anscombe,
1979b, p. 211).

A final comment. Rosalind Hursthouse writes:

I have found that when … discussing papers by … Anscombe … with
fellow philosophers, that what often blocks understanding is the
unconscious assumption that everyone shares the view that, for exam-
ple, beliefs and desires are natural kinds, or that a reason is a belief/
desire pair that causes an action, or that all mental states or brain
states – or, more generally, that philosophy is supposed to uncover
or construct the foundations of our thought. […] Sometimes – not
always, of course, the cloud lifts if one says ‘But you don’t believe that
so-and-so if you’re a Wittgensteinian’.

(Hursthouse, 1999, p. 16)

Hursthouse is right that these and many others are things that
‘you don’t believe if you’re a Wittgensteinian’ and that those are
the very things that Anscombe doesn’t believe. It is also absolutely
true that these ‘unconscious assumptions’ will block understand­
ing of Anscombe. Of course, when these assumptions are made
conscious one might chose not to reject them – certainly the fact
that a Wittgensteinian would do so is no argument. However,
Anscombe did and, in many places in her work, she offered
sophisticated arguments why one ought. Some of the best of them
are in Intention.

NOTES

1 For an updated bibliography of Anscombe’s published writings, see
www.unav.es/filosofia/jmtorralba/anscombe_bibliography.htm
(accessed 8 September 2015).

2 This error was compounded, if not caused, by Goldman’s coining of
the label ‘Davidson-Anscombe view’ for a thesis about action identity
(Goldman, 1970, p. 1). Much of what was written on Anscombe’s Intention
through the 1970s and 1980s was in fact concerned with defending or

http://www.unav.es/filosofia/jmtorralba/anscombe_bibliography.htm

INTRODUCTION 9

repudiating this thesis, a thesis which does not in fact appear in any of
Anscombe’s writing.

3 Future citations will have the form section number + page number
(e.g. §1, p. 1). All references are to the second impression of the sec-
ond edition, published in 1957. This is the version standardly referred to
(copies of the first edition, and the first impression of the second are hard
to come by). As Anscombe notes in her introduction, some minor altera-
tions were made between the first and second edition (§§19, 33 and 34);
and the first and second impression of the second edition (in §§2, 6, 17,
33 and 34). Those with particular interest in these paragraphs may wish
to consult earlier editions.

4 This discussion owes much to Cora Diamond’s analysis of the rela-
tion between Frege’s and Wittgenstein’s philosophical method, in
‘Introduction I’ to her The Realistic Spirit (1991).

5 This characterisation of the way in which the Fregean idea manifests in
Wittgenstein’s later work is from Cora Diamond (1991), esp., pp. 4–6.

6 Roger Teichmann (especially 2014a, 2014b) and Cora Diamond (1991,
1966, 1997) are notable exceptions.

7 Those familiar with recent work on Anscombe may be shocked by the
absence of any discussion of ‘the progressive’ in elucidating the notion
of practical knowledge or of ‘direction of fit’ in relation to ‘Theophastus’
Principle’. Anscombe does not use these terms, and I have not found
them helpful in presenting her arguments. I leave it to others to make a
case for their insertion.

1
ELIZABETH ANSCOMBE

AND INTENTION

Gertrude Elizabeth Mary Anscombe was born in 1919 in Limerick,
Ireland, where her father had been posted during the First World
War. Her family soon moved to the south of England, where she
attended Sydenham High School. At the age of twelve Anscombe
discovered Roman Catholicism when reading about the persecution
of priests in England under Queen Elizabeth I, and on beginning
at Oxford University in 1938 she took formal instruction in the
religion with Father Richard Kehoe, a Dominican priest. She was
introduced to the philosopher Peter Geach, a fellow convert. They
were married in 1941 and went on to have seven children.

As an undergraduate at Oxford, Anscombe studied Classics
(or ‘Mods and Greats’) – a combination of ancient Greek and
Roman philosophy, history and literature. At the insistence of her
philosophy tutors Anscombe was awarded First Class Honours
despite spectacularly failing the Roman History paper. Famously,
when asked in her viva ‘Miss Anscombe, is there any fact at all
about the history of Rome which you would like to comment on?’
she replied ‘No’ (Teichman, 2002, p. 31).

ELIZABETH ANSCOMBE AND INTENTION 11

Anscombe moved to Cambridge as a research student in philos-
ophy between 1941 and 1944. In 1946 she returned to Oxford, to
Somerville College, taking up a Research Fellowship during which
she wrote Intention. Between 1970 and her retirement in 1986,
she held the chair at Cambridge that had been Wittgenstein’s. She
died in 2001, aged 81.

ANSCOMBE’S TIMES

When Anscombe began her undergraduate studies at Oxford in
1938, the University had been admitting women for only 50 years,
and awarding them degrees for fewer than 20. Oxford during this
time seems to have been characterised by a suspicious and vaguely
hostile attitude toward the women students. Iris Murdoch, on
arriving at Oxford in the same year as Anscombe, was warned
by her tutor to ‘be careful how you behave … [T]he women are
still very much on probation at this University’ (Conradi, 2001,
p. 82). In ‘Mr Truman’s Degree’ – written 20 years later – Anscombe
describes how the men of St John’s college were convinced to block
her opposition to the degree: ‘The dons at St John’s were simply
told “The women are up to something in the Convocation: we have
to go and vote them down”’ (Anscombe, 1957c, p. 65). Writing in
the late 1950s, the journalist Mehta described the women’s ‘inva-
sion of the field’ (Mehta, 1965, p. 27) that had taken place during
the war – an exaggerated metaphor which does not suggest the field
was altogether glad of them.

With the Second World War on the horizon, Military Service
was introduced in 1939, with conscription for men aged 18 to 41
coming the following year. Although some academics were reserved,
many were called up, along with numerous of Anscombe’s con-
temporaries. By 1940 the women found themselves outnumbered
only 2:1 at Oxford, as compared to 5:1 in 1936. The humanities
were hit particularly hard by conscription, as, unlike medicine or
science, there was no prima facie case for thinking that their studies
were making a contribution to the war effort. This meant that far
fewer were reserved.

This context provided the women students with a unique
climate in which to develop. As well as Anscombe, this cohort

ELIZABETH ANSCOMBE AND INTENTION12

included Philippa Foot, Iris Murdoch, Mary Midgely and Mary
Warnock, all of whom became brilliant philosophers. It is surely
no coincidence that so many of the wartime group went on not
only to be influential intellectuals but also to write and think in
ways that stand quite outside fashionable or mainstream style.

Mary Midgley recently described this ‘Golden Age of female
philosophy’:

As a survivor from the wartime group, I can only say: sorry, but the
reason [why so many well-known female philosophers emerged
from Oxford soon after the war] was indeed that there were fewer
men about then. The trouble is not, of course, men as such – men
have done good enough philosophy in the past. What is wrong is a
particular style of philosophising that results from encouraging a lot
of clever young men to compete in winning arguments. These people
then quickly build up a set of games out of simple oppositions and
elaborate them until, in the end, nobody else can see what they are
talking about. […]

It was clear that we were all more interested in understanding this
deeply puzzling world than in putting each other down. That was how
Elizabeth Anscombe, Philippa Foot, Iris Murdoch, Mary Warnock and
I, in our various ways, all came to think out alternatives to the brash,
unreal style of philosophising – based essentially on logical positivism –
that was current at the time. And these were the ideas that we later
expressed in our own writings.

(Midgley, 2013)

While it is important not to underestimate the differences between
these women’s philosophical views – for one thing, Anscombe’s
Catholicism set her apart from the others – what undeniably unites
these women’s work is an ethical outlook, according to which the
way to proceed is to study not just man but man in his circum-
stances; not just individual moral agents but the moral environment
that we share. Anscombe, echoing Wittgenstein, wrote that ‘[i]t
belongs to the natural history of man that he has a moral envi-
ronment’ (Anscombe, undated b, p. 224). Iris Murdoch described
herself, Foot and Anscombe as united in their rejection of the idea
that ‘the human being was the monarch of the universe, that he

ELIZABETH ANSCOMBE AND INTENTION 13

constructed his values from scratch’ and in their shared interest
in ‘the reality that surrounds man – transcendent or whatever’
(Mehta, 1965, p. 52).

As well as providing freedom from the kind of stultifying
‘games of simple opposition’, and creating a community of female
philosophers who continued to write and think together through-
out their lives, the war was part of a ‘deeply puzzling world’ with
which to get to grips, and lent real urgency and seriousness to
that task. Across the globe the competing ideologies of Nazism,
Communism and Liberalism were being realised in societies then
pitted against one another in war; human beings were engaged
in methodical and bureaucratised mass slaughter, torture and
destruction.

When the war ended in 1945, those of Anscombe’s male col-
leagues who had survived returned from service. Stuart Hampshire
and J. L. Austin returned from interrogating enemy prisoners on
behalf of the British Intelligence Corps; R. M. Hare from three
years in a Japanese Prisoner of War camp; P. F. Strawson from
Italy where he served with the Royal Electrical and Mechanical
Engineers. For some, these experiences had a profound effect on
their philosophy. Hare’s moral theory, for example, was famously
shaped by his experiences in Japan and his realisation – in perfect
opposition to the ethical perspective described by Murdoch – that
‘nothing was “given” in a society, that … every man was born with
a conscience, and this, rather than anything in society … was the
source of morality’ (Mehta, 1965, p. 48).

OXFORD PHILOSOPHY

Against this world-historical background, a major shift was
playing out in Western philosophy. This shift is known as the
‘linguistic turn’. At the beginning of the twentieth century, Frege
and Wittgenstein, in different ways, made the case for thinking
that the way to answer philosophical questions might be to study
words and concepts rather than things and essences.1 Their idea –
recognised in the coming decades as revolutionary – was that the
need for elaborate metaphysics to answer to philosophical ques-
tions about the nature of, for example, being and substance and

ELIZABETH ANSCOMBE AND INTENTION14

mind, could be eliminated if careful attention was paid to language
and to its logical form. To give a crude example: rather than asking
‘What is Truth?’ we should ask, ‘What does “Truth” mean?’. The
claim they made – and sought to demonstrate in the Begriffsschrift
(Frege, 1879) and Tractatus (Wittgenstein, 1922) – was that phil-
osophical problems will simply disappear, once we are clear about
the grammar of the language in which they arise.

With this methodological turn arose a new conception of the
kind of activity in which a philosopher is engaged. If the method
were correct then the philosopher’s task was more like that of
the mathematician than the empirical scientist (c.f. Frege, 1956,
p. 308). His or her aim was to uncover and describe the formal
structure of language and not to make new discoveries about the
nature of Reality, as previous generations of philosophers had
attempted. Indeed, this method was to be a prophylactic against
the urge to build baroque ontological systems.

Wittgenstein captured this change of method and self-conception
when he wrote in the Tractatus:

The object of philosophy is the logical clarification of thoughts.
Philosophy is not a theory but an activity. The result of philosophy is
not a number of ‘philosophical propositions’, but to make proposi-
tions clear.

(Wittgenstein, 1922, 4.112)

The question of how one implements the method of linguistic
philosophy – how one undertakes Frege and Wittgenstein’s task of
‘the logical clarification of thoughts’ – depends to a large degree
on what one thinks about ordinary language, that is, English or
French or Chinese, as it is spoken outside the philosophy class-
room. For some philosophers – for example Frege, Carnap and the
early Wittgenstein – ordinary language was irredeemably vague,
ambiguous and ill-formed. It was thus inadequate to the task of
expressing thoughts and propositions clearly and univocally. The
linguistic philosopher’s task was thus to devise a formal or ideal
language in which thoughts and propositions could be adequately
expressed. Such a language, it was thought, would not be one in
which philosophical problems could arise.

ELIZABETH ANSCOMBE AND INTENTION 15

For a second group of philosophers, ordinary language was
perfectly precise and wholly adequate to its task. Problems arose
only when philosophers tried to use ordinary language in unor-
dinary ways. For example, instead of asking ‘What is the time?’
or ‘Do you have time to see me now?’ – questions in perfectly
good standing – metaphysicians ask ‘What is Time?’. In doing so,
they misuse ordinary language. Linguistic philosophers belonging
to this second group – centred on J. L. Austin – are known as
the ‘ordinary language philosophers’ and were in their heyday in
Oxford during the 1940s and 1950s. They proceeded ‘by exam-
ining what we should say when, and so why and what we should
mean by it’ (Austin, 1957, p. 8). Austin described their resources
as ‘the dictionary’, ‘the law’, and ‘psychology’, the latter including
‘anthropology and animal behaviour’ (Austin, 1957, p. 14). These
philosophers are less mathematicians than ethnographers. The
fact that it would be unnatural to use a word in a context in which
a metaphysician is inclined to use it, could be a crucial move in
revealing the source of a particular philosophical error. For exam-
ple, in ‘A Plea for Excuses’, Austin argues that the problem of free
will can be solved by ‘examining all the ways in which each action
may not be “free”, i.e., the cases in which it will not do to say
simply “X did A”’ (Austin, 1957, p. 6). Once this examination is
complete, we will see that the case that the metaphysician wants
to insist on simply cannot be described.

Austin’s account of involuntary action is dismissed by
Anscombe in Intention (§7, p. 12), and so there is an immediate
question to be asked about whether she should be classified as an
‘ordinary language philosopher’, alongside her Oxford contempo-
raries. Anscombe was certainly a linguistic philosopher, but her
way of understanding the methods of linguistic philosophy was
shaped more by Wittgenstein’s than by Austin’s. From the 1940s,
Anscombe travelled from Oxford to Cambridge regularly to
study with Wittgenstein, and the pair became friends. She trans-
lated his Philosophical Investigations from German into English,
and her translation has not been surpassed. The influence of the
Investigations and of Wittgenstein’s way of approaching philosophi-
cal questions is apparent throughout Anscombe’s work and no more
so than in Intention. Superficially: Intention – like the Philosophical

ELIZABETH ANSCOMBE AND INTENTION16

Investigations and unlike Anscombe’s other writings – is arranged
as a series of numbered paragraphs, and the Investigations is one
of only a handful of texts to which she refers. But the influence is
much deeper than that. The book, as we will see, is an exemplifica-
tion of the method of grammatical investigation.

In Chapter 2 we will examine that method in some detail.
Here, we can place a marker by noting a fundamental difference
between Wittgenstein’s method of grammatical investigation and
Austin’s study of ordinary language. The difference lies in the
role each assigns to what we say about language rather that in
language. For Wittgenstein, what we say about language – our
linguistic intuitions if you like – are not linguistic data for the
philosopher to accommodate, but are themselves a potential
source of confusion.2 The fact that, for example, we do not find it
‘natural’ to apply the word ‘voluntary’ to acts like raising a hand
or walking across the road – the fact that we use it most naturally
when we are interested in excuse and blame – is itself a datum
that might lead us to mistake the grammar of that concept. In the
next chapter we will consider what the data for a grammatical
investigation might then be.

By the time Anscombe was writing Intention a ‘post-Linguistic
thaw’ (Strawson, 1960, p. 584) was underway and the method
of linguistic philosophy was beginning to seem inadequate to
the task of doing deep and creative philosophical work. Mehta
paints a portrait of a strange intellectual climate, generated by
a self-consciously negative methodology and a nostalgia for the
days when Oxford was preparing young men to govern Empire
(1965, p. 37). The philosopher and social anthropologist Ernest
Gellner – whose view we must take with a pinch of salt given
his famous opposition to the Oxford philosophers3 – describes
Oxford as ‘a nursery for leaving the world exactly as it is’ (Mehta,
1965, p. 39) while Bertrand Russell explains that ‘present-day
Oxford philosophy is gentlemanly in [the] sense [that] it takes
nothing seriously’ (Mehta, 1965, p. 42).

By the standards of Oxford philosophy, Anscombe was no gentle-
man. As we will see in the next chapter, she took things extremely

ELIZABETH ANSCOMBE AND INTENTION 17

seriously and had no desire to leave the world exactly as it is. She
was fiercely critical of her colleagues, her University, her discipline,
her Church and ultimately of what goes for ‘the highest and best
ideals of the country at large’ (Anscombe, 1957b, p. 163). This
opposition regularly attracted fierce rebuke from her colleagues,
who objected not least to her willingness to label them – or at least,
their philosophy – ‘corrupt’, ‘corrupting’, ‘stupid’, ‘ murderous’,
and more.4 This is not the kind of language one usually finds in a
philosophical discussion. Much of what she rails against is – what
she sees as – a refusal to think carefully, hard and seriously about
extremely serious things. About life and death and torture and
war and sex and the moral education of children. Anscombe is
not an elitist; for her, philosophy is not primarily a technical or
scholarly discipline but simply ‘thinking about the most difficult
and ultimate questions’, something she found young children to
be capable of (Geach, 2005, p. xiii). It goes with this expectation
that when clever people fail or refuse to think hard, clearly, or
carefully, she views them with contempt. This is why her most
forthright criticism is reserved for those who speak from positions
of authority and education.

Reflecting on her paper ‘War and Murder’ she described it as
‘written in a tone of righteous fury about what passed for think-
ing about the destruction of civilian populations’ (ERP, p. vii).
Jenny Teichman describes Anscombe’s tone in that paper as ‘like
the Prophet Jeremiah’s’ (Teichman, 2002, p. 49). Many would
say (wrongly in my view) that righteous fury has no place in the
urbane setting of a philosophy seminar.

ANSCOMBE’S FAITH

That philosophy ought not to take anything seriously was not the
only gentlemen’s agreement in place in Oxford during the 1950s.
The other – connected I suppose – was that philosophy ought to
be conducted in a secular manner. To speak in the broadest terms,
while Medieval and Early Modern philosophy is characterised by
contemplation of God, and an attempt to employ human rea-
son to reach a – necessarily inadequate – understanding of His
powers and attributes, post-Enlightenment philosophy replaced

ELIZABETH ANSCOMBE AND INTENTION18

God with Man as the object of contemplation. The preoccupa-
tion of twentieth- and twenty-first-century philosophy was, and
is, human nature, human powers and attributes. According to this
second gentleman’s agreement, if a philosopher is foolish enough
to be a theist, this is a private matter and should be kept out of
her philosophy.5

For philosophers or theologians who are familiar with the
writings of St Thomas Aquinas, the account of action set out
in Intention will be extremely familiar. Much of the best recent
work on Anscombe has been concerned to illuminate the deeply
Thomistic character of Intention, and of her moral philosophy
more widely. Aquinas – a thirteenth-century Dominican friar,
canonised 50 years after his death – is recognised as the Catholic
Church’s greatest philosopher and theologian. His philosophical
works demonstrated in detail how Aristotle’s teaching – newly
available to medieval scholars – could be integrated with Catholic
doctrine. In the introduction I followed Hursthouse in sug-
gesting that ‘But you don’t believe that so-and-so if you’re a
Wittgensteinian’ is a useful reminder. So too is ‘Read Anscombe
in light of Aquinas’ (Schwenkler, 2015, p. 10). Viewed from one
perspective, Intention is an elucidation and defence of Aquinas’s
practical philosophy, and provides the foundation for Catholic
morality and the doctrine of double effect.

The assumption that religious belief is somehow incom-
patible with linguistic philosophy may derive in part from the
anti-metaphysical bent of that movement, and may go some
way toward explaining why Anscombe did not make this influ-
ence more explicit. But for serious philosophers wishing to put
forward Thomistic ideas, a further barrier to charitable hearing
was a prevalent view of Aquinas’s work as deeply unphilosophical.
Pope Leo XIII’s 1879 encyclical Aeterni Patris, made Aquinas the
‘official’ philosopher of the Catholic Church, and this led to the
production of numerous textbook works, designed for teaching
Aquinas’s philosophy to Catholics. Writers of these textbooks
were less concerned with philosophical argument and nuance,
than with doctrine and summary, and so the picture of Aquinas
that emerged in them was of an uncritical and dogmatic thinker.
Mary Geach, Anscombe’s daughter, recalls: ‘she said to me that it

ELIZABETH ANSCOMBE AND INTENTION 19

aroused prejudice in people to tell them that a thought came from
him: to my sister she said that to ascribe a thought to him made
people boringly ignore the philosophical interest of it, whether
they were for Aquinas or against him’ (Geach 2011, p. xix).

Anscombe’s view of her faith plays an important role in shaping
the content of her moral philosophy and psychology. She rejects a
conception of Christianity as ‘an ideal and beautiful religion, imprac-
ticable except for a few rare characters’, which ‘preaches the God
of love whom there is no reason to fear’ (Anscombe, 1961, p. 55).
To this conception she contrasts ‘the truth about Christianity’: it
is ‘practicable but severe’ (p. 56). These go together: the severity
requires the practicability because it would be unjust to punish those
who failed to live up to standards achievable only by saints.

Anscombe thought that it must be possible for everyone to
avoid doing evil. This provides the context for her fierce attacks
on pacifism. She thought that the loftiness of the standards that
pacifism – and the beautiful but impracticable idea it embodied –
imposed made it not merely false but positively dangerous. It
leads, she says, to a ‘nonsensical “hypocrisy of the ideal stand-
ard”’ (Anscombe, 1957c, p. 69). By drawing the line between
what is and is not permitted in such a place that one is always
already doing wrong, the pacifist encourages the thought that
‘we have to accept the evil’, which then leads to the idea that
‘once you are in for it, you have to go the whole hog’ (Anscombe,
1957c, pp. 69–70).

We will see this question about the drawing of boundaries
again in Chapter 2. One of the fundamental tasks of Intention
is to provide a philosophy of psychology which can ground the
‘truth about Christianity’. Questions Anscombe later framed and
answered using the philosophical account she set out in Intention
include: When is an act of killing murder (1961)? When is caus-
ing the death of a foetus abortion (1968)? What are the proper
limits of a state’s coercion of its citizens (1978)? How should one
provide a moral education for one’s children (1962a)? Is capital
punishment acceptable (1957c)? Should control groups be used
in clinical trials (1983b)? How should severely disabled children
be cared for (1981)?

ELIZABETH ANSCOMBE AND INTENTION20

When we turn to look at Anscombe’s answers to those ques-
tions, it becomes clear that Anscombe drew the boundaries in
line with traditional Catholic teaching. For many readers, in
particular those who do not share her faith, Anscombe’s answers
to particular moral questions are unacceptable: abortion, con-
traception, gender reassignment surgery, in vitro fertilisation
treatment, voluntary euthanasia, and sex outside heterosexual
marriage all come out as prohibited; capital punishment, war and
state coercion do not.6

However much one dislikes where she draws the line, what she
is attempting is to put it in a place which makes it possible for
ordinary people – and not just moral saints – to be good and to be
virtuous, and which means it is no longer open to say ‘I couldn’t
do it, it’s not realistic’: yes, you could, and it is. The alternatives
are to draw no line, preferring instead to weigh every case in light
of the circumstances; or to draw a line which means we just have
to accept we’re all compromised and make the best of it on pain
of hypocrisy.

INTENTION’S INFLUENCE

Anscombe’s monograph has been extremely influential but,
somewhat paradoxically, the account of intention it contains has –
at least until very recently – made very little impact on the way
in which philosophers think about action and intention. This is
because it has been so little understood. The initial reception of
Intention reflects this ambiguity. Reviews were published in at
least seven philosophy journals and in the Catholic periodical,
The Tablet, and opinion on the significance of the work was
fairly unified. While few agreed with Judith Jarvis’s assessment
that ‘Miss Anscombe’s book is easily one of the best things to
come from England in many years’ (Jarvis, 1959, p. 31), most
recognised the importance of Anscombe’s treatment of her topic
and expressed confidence that Intention would become a classic
text. However, recognition of the book’s significance did not
go hand-in-hand with a grasp of the book’s thesis. P. L. Heath
expressed frustration that ‘the work seems hardly to have got
beyond the stage of a first draft’ (Heath, 1960, p. 281) and

ELIZABETH ANSCOMBE AND INTENTION 21

Joseph Cowen, while admiring Anscombe’s ambition, acknowl-
edged that the book was likely to induce a ‘general feeling of
malaise’ in many readers (Cowen, 1959, p. 146). The reviewer
in The Tablet remained unaware of the book’s Thomistic core,
and while recommending it for providing ‘mental exercise of a
kind which is most valuable in virtue of its very strenuousness’,
worried that the results were ‘not very informative’ and on the
whole ‘banal’ (Trethovvan, 1958, p. 12). Philosophers writing
in the coming decades remained confident of the book’s impor-
tance, but still the account of intention Anscombe offers in it
made almost no impact.

One reason why philosophers have been confident of the sig-
nificance of Intention – despite failing to understand its central
insights – is its status as the genesis of its subject area: it estab-
lished action as a philosophical topic in its own right. Prior to
Intention, one had to go back to Thomas Reid’s Essays on the Active
Powers of Man, published in 1788, for a book-length treatment
of the subject in English. This means that, even while Intention
has remained misunderstood, it has set the agenda for philoso-
phy of action. Three questions, seemingly posed and answered by
Intention, came to predominate the discussion during the 1960s
and 1970s, and continue to shape discussion today.

First: what are the identity conditions for actions? Or, how do we
count actions? In a now famous discussion, Anscombe describes a
man who is moving his arm in order to operate a pump, where the
pump is replenishing the water supply of a house with poisoned
water. The man’s intention is to poison the inhabitants. Anscombe
asks: ‘Are we to say that the man … is performing four actions?
Or only one?’ (§26, p. 45).

As is shown by Anscombe’s two proposed answers, this ques-
tion sets up a debate about whether it is bodily actions or intentions
that provide the identity conditions for actions. This man is per-
forming one bodily action, but has four intentions: how many
intentional actions does that give us? The so-called ‘Identity Thesis’
is the view that it is bodily actions that matter. Anscombe herself
says that the ‘only distinct action of his that is in question is …
moving his arm up and down with his fingers round the pump
handle’ (§26, p. 46). So, Intention is seen as the genesis of the

ELIZABETH ANSCOMBE AND INTENTION22

Identity Thesis, a thesis that seems to depend on an account of
event identity that Anscombe neither articulates nor defends.

The second question is: what is the relation between reasons
for action and causes of action? Or, to put it another way, are
mental states like intention, belief and desire the causal ante-
cedents of actions? Intention argues that answers to the question
‘Why?’ which give an action’s cause, must be distinguished from
those that give a reason for acting (e.g. §§9–10). Following
Intention, a ‘Wittgensteinian school’ emerged in the philosophy
of action. These philosophers cast Anscombe’s remark about
the distinction between reasons and causes in terms of a gen-
eral thesis about causation: if the connection between a and b is
conceptual then a cannot cause b. So, they argued, because the
connection between intention and action is conceptual it cannot
be causal. Nor, as such, could it be understood in terms available
to the natural sciences.

The third question is: what is the connection between my
intention to act, my intentional action, and the reason that justi-
fies or explains that action? Anscombe, in introducing her topic
under three heads (§1), seemed to pose this challenge, meeting
which was recently described by Kieran Setiya as ‘the principal
task of the philosophy of intention’ (Setiya, 2014). Anscombe’s
Intention, for reasons we will discuss in the next chapter, has come
to be associated with answers to this question which treat the
application of the concept of intention to actions as fundamental.

In 1963, Donald Davidson published what has become an
extraordinarily influential paper: ‘Actions, Reasons and Causes’.
In that paper, and a series of later papers on action and intention,
Davidson defended the Identity Thesis, while at the same time
repudiating the claim that the conceptual connection between
desires and actions ruled out a causal connection.7 Despite reject-
ing the separation of reason and cause, he claimed to have derived
many of his central insights from Anscombe’s Intention. Indeed,
he seemed to be offering a way to acknowledge what Anscombe
was on to when she insisted on that separation, and to retain a
role for the question ‘Why?’ in eliciting reason-giving explana-
tion, while also allowing a causal-role for belief and desire in
action. As Davidson’s account of action acquired the status of the

ELIZABETH ANSCOMBE AND INTENTION 23

‘standard story’, Intention came to be seen as foreshadowing that
account. The Identity Thesis came to be known as the ‘Davidson-
Anscombe thesis’, and the deeply anti-Humean character of
Anscombe’s thought became obscured.

In 1959, P. F. Strawson wrote of the ‘post-linguistic thaw’. He
identified Intention as a part of the new scholarship which he saw
as transforming linguistic philosophy into something that was pre-
pared to move beyond ‘piecemeal studies’ toward ‘broader synthesis’,
and which was not afraid to recover the grand metaphysical systems
of their philosophical predecessors (1960, p. 2515). In the last dec-
ade, philosophers seeking to precipitate a similar thawing-effect
on contemporary analytic philosophy have returned to Intention,
and to Anscombe’s work more widely, in search of tools to do so.
Disillusioned with the reductive, physicalist programme that has
come to be predominant in analytic philosophy of mind, philoso-
phers returning to Anscombe have looked for a key to unlock the
insights of pre-Modern thinkers – like Aristotle and Aquinas – for
modern analytic philosophy. ‘Analytic Aristotelianism’ – associated
in particular with Michael Thompson – and ‘Analytic Thomism’ –
associated with John Haldane – are two such movements, each
heavily indebted to Anscombe’s work. In this second epoch there
has been a serious attempt by philosophers of action to return to
Intention and engage with Anscombe directly. These philosophers
have begun to uncover an account of action that is deeply anti-
thetical not only to contemporary reductionist programmes in the
philosophy of mind, but also to the prevailing Davidsonian ortho-
doxy in action-theory, to neo- Humean philosophy of psychology,
and to Cartesian metaphysics. It is perhaps too early to judge the
significance of this movement for understanding and appreciating
Anscombe’s work.

Despite the wide agreement that Intention is highly significant,
and the work that I have just described, Intention – though often
quoted and lauded – is still not widely read. The problem is that
the book is so difficult. Kurt Baier, an earlier reviewer, recog-
nised it as a ‘highly stimulating, penetrating, and original book’
(Baier, 1960, p. 81) but admitted that he would ‘probably not

ELIZABETH ANSCOMBE AND INTENTION24

have persevered in looking for [the book’s over-all theme] if [he]
had not undertaken to write [a] review’ (p. 71). Writing this
Guidebook, I can well empathise.

NOTES

1 This slogan is from Ernest Gellner’s Words and Things, published in 1959.
2 For a clear and persuasive exposition of the difference between the idea

of a grammatical investigation as it occurs in ordinary language philos-
ophy and as it occurs in Wittgenstein’s later work, see Gordon Baker
(2001). Readers may also find this useful in illuminating aspects of
Anscombe’s method in Intention.

3 His book Words and Things led to a major incident when Gilbert Ryle,
heavily criticised by Gellner, refused to have the book reviewed in the
philosophy journal Mind, of which he was at the time editor. Bertrand
Russell stepped into the fray and a series of letters in The Times followed.

4 See, for example, the exchange of letters in The Listener, following
the publication of the transcript of ‘Does Modern Moral Philosophy
Corrupt the Youth?’ (R. M. Hare and P. H. Nowell Smith 1957, p. 311 and
Anscombe 1957d, p. 349).

5 Mary Geach describes this attitude, prevalent still, in her introduction
to FHG. ‘Some people’, she reports, ‘have found it surprising that
Peter Geach and Elizabeth Anscombe (my parents) should have been
distinguished members of the analytic school, while at the same time
practising and believing the Catholic religion’ (p. xiii).

6 Of course, to say that a certain kind of thing – war, say – is not outright
prohibited is not to say that it is permitted in any particular case. The
relationship between prohibition and permission is not symmetric in
this way. For an example of this, see Anscombe (1939).

7 For a helpful discussion, see Stoutland (2011), ‘Introduction’, sections 2
and 3, pp. 6–12.

SUGGESTED READING

Anscombe, G. E. M. (1976a). ‘The Question of Linguistic Idealism’. From ‘Essays on
Wittgenstein in honour of G. H. von Wright’, Acta Philosophica Fennica, 28, 1–3.
Reprinted in FPW, pp. 112–133.

Haldane, John (2004). ‘Thomism and the Future of Catholic Philosophy’. In Faithful
Reason: Essays Catholic and Philosophical (London: Routledge).

ELIZABETH ANSCOMBE AND INTENTION 25

Rorty, Richard (1967). ‘Metaphilosophical Difficulties of Linguistic Philosophy’. In
The Linguistic Turn: Recent Essays in Philosophical Method (Chicago and London:
University of Chicago Press), pp. 1–39.

Stoutland, Frederick (2011). ‘Anscombe’s Intention in Context’. In Anton Ford,
Jennifer Hornsby and Frederick Stoutland (eds), Essays on Anscombe’s Intention
(Cambridge, Mass.: Harvard University Press).

Teichman, Jenny (2002). ‘Gertrude Elizabeth Margaret Anscombe: 1919–2001’,
Biographical Memoirs of Fellows, Vol. 115 (Oxford: Oxford University Press),
pp. 30–50.

2
THREE TASKS FOR

INTENTION

The difficulties posed for the reader by Anscombe’s little book are
manifold, but the first is how to begin it. It is impossible not to
be mystified on first reading Intention. Anscombe’s book contains
neither a recognisable introduction, nor conclusion. There are no
chapters, but rather 52 paragraphs of varying length and – one
supposes – varying importance. It is not clear that there is a single
line of reasoning, nor even what kind of argument is being put
forward. The book seems utterly to resist summary or paraphrase,
and the table of analytic contents which precede §1 are useful for
navigation but do not appear to offer much help beyond that.

One way to approach the question of how to begin Intention is
to look at how Anscombe began it. Anscombe wrote the lectures
which became the book after a specific event: Oxford University
proposed to award an honorary degree to the former president of
the United States, Harry S. Truman. At the time of the award,
Anscombe had been covering Philippa Foot’s teaching on an
undergraduate module on ethics. She was dismayed by what she
discovered about the state of contemporary moral philosophy; for

THREE TASKS FOR INTENTION 27

reasons we will see below, she thought it to be corrupt and shallow.
To her horror, she then encountered this same mode of moral
thinking being deployed to defend Truman’s wartime bombing of
Japan. These circumstances combined to make it clear to her that
‘there were some things which she understood and [her contem-
poraries] did not’ (Geach, HLAE, p. xiv).

Anscombe came at this time to think that ‘it is not profitable …
to do moral philosophy … until we have an adequate philosophy of
psychology’ (Anscombe, 1958b, p. 169), and she identified three
pieces of conceptual clarification which were necessary if work in
ethics were to be possible. First, an account would be needed of
the distinction between an action’s intended and its merely foreseen
consequences. Second, a criterion for the intention with which a
person acted would be required. This criterion must show that and
why a mere ‘act of will’ on the part of the agent is not sufficient
to determine the intention with which an action is done. Third,
ethics would require an account of human action – voluntary and
intentional – which could make sense of the idea that actions can
instantiate virtues (that is, that a particular act, say, my handing
you a piece of paper, can be an act of kindness, or justice, or honesty).
These three tasks, as anyone familiar with Intention will recognise,
correspond loosely to Anscombe’s treatment of her topic under
its ‘three head’: ‘expression of intention for the future, intentional
action, and intention in acting’ (contents, §1, p. i).

For pre-modern philosophers like Aristotle and Aquinas, the
ethical import of Intention – and need for an adequate philosophy
of psychology as a preliminary to moral philosophy – would have
been immediately obvious; but from a Modern – post-Cartesian –
perspective, it seems anything but. It is natural for us, now, to
suppose that the questions that concern ethics – normative ques-
tions about right and wrong and about how one ought to live
and metaethical questions about the nature of ethical properties –
can be answered quite independently of determining the nature
of intention or intentional action, topics that belong to the phi-
losophy of mind or psychology. In this chapter we will look at
Anscombe’s claim that ethics ungrounded in an understanding of
intention is necessarily corrupt. This will enable us to outline the
task of Intention.

THREE TASKS FOR INTENTION28

THE GENESIS OF ANSCOMBE’S INTENTION

Harry S. Truman became president of the United States in April
1945 when President Franklin D. Roosevelt died in office. This
was almost three and half years after Roosevelt entered the country
into the Second World War, on the side of the Allies, following the
bombing of Pearl Harbor. Truman was President for the remainder
of hostilities, and it was on his order that atomic bombs were
dropped on Hiroshima and Nagasaki in 1945. The bombs killed
between 75,000 and 125,000 civilians on impact, with the same
number again dying in the next few months.

In 1956, Oxford University proposed to award Truman with an
honorary degree. Anscombe opposed this on the grounds that, as
the person who gave that order, Truman was guilty of mass murder.
As such, she argued, Oxford University should not be bestowing
him with honours; to do so, she said, was comparable to ‘honour-
ing Hitler or Nero or Genghis Khan’ (Geach, HLAE, p. xiv).

The putative justification for the dropping of the bombs was –
and still is – that ‘it was right to massacre the Japanese because it
was … productive of a better total state of affairs than not doing
so would have been’ (Anscombe, 1957b, p. 164). The deaths of all
those people was the least bad outcome given the way in which the
war with Japan was unfolding. There was very good evidence that
the Japanese army was prepared to continue to fight beyond the
point at which an army might usually be expected to capitulate
and that their soldiers would choose suicide over surrender. This
meant that a land-invasion, were it attempted by the Allies, would
be likely to lead to very many lives being lost on both sides, and
many atrocities being visited upon an already beleaguered civilian
population. The Americans’ experience at the Battle of Okinawa
was taken as a warning. The land-invasion of the Japanese island
led to a battle lasting 82 days in which 90,000 soldiers (US and
Japanese) and up to 150,000 civilians were killed. Mass suicides at
the prospect of capture were reported among the Japanese people.
Faced with the terrifying spectacle of an army who would fight
without care for their own lives, and a civilian population who
were prepared to endure such horror and loss, Truman ruled out a
land-invasion in order to ‘prevent […] an Okinawa from one end
of Japan to the other’ (Kennedy, 1999, p. 410).

THREE TASKS FOR INTENTION 29

An alternative to land-invasion presented itself in the form
of the atomic bomb and on 26 July 1945 the Potsdam Decla-
ration, signed by Truman, Winston Churchill (the British
Prime Minister) and Chiang Kai-shek (the Chairman of the
Nationalist Government of China), was issued. The Declaration
had the form of an ultimatum. It stated that unless the Japanese
surrender unconditionally, they would face ‘prompt and utter
destruction’. When no surrender came, an atomic bomb was
dropped on Hiroshima on 6 August. Three days later, with
Japanese surrender still not forthcoming, a second bomb was
dropped on Nagasaki. On 2 September the Japanese Emperor
issued an unconditional surrender.

What should we think of the men who signed the Potsdam
Declaration and who delivered on the ultimatum by ordering
atomic bombs to be dropped on two Japanese cities, bombs
that were designed indiscriminately to kill or maim – vaporise,
burn, poison, and tear to pieces – a population of unarmed
men, women, and children? One common response is that they
did what had to be done. They took an awesome decision in
the face of a terrible set of choices. Looked at in this light such
a decision might be seen (somewhat grotesquely given it was a
decision to burn to death some babies) as a kind of heroism:
Truman was prepared to do something morally repugnant – to
get his hands dirty – for the greater good. As Anscombe puts it,
people seem to think that ‘Mr Truman was brave because … what
he did was so bad’ (Anscombe, 1957c, p. 64). Seen this way,
to not drop the bomb might look like cowardice, self-interest,
vanity, or the avoidance of duty, and opposition to it a kind of
‘high-mindedness’. Anscombe puts this point in the mouth of
an interlocutor:

The action was necessary …; it probably saved more lives than it sac-
rificed; it had a good result, it ended the war. Come now: if you had to
choose between boiling one baby and letting some frightful disaster
befall a thousand people – or a million people, if a thousand is not
enough – what would you do? Are you going to strike an attitude and
say ‘You may not do evil that good may come’?

(Anscombe, 1957c, p. 65)1

THREE TASKS FOR INTENTION30

Anscombe’s opposition did not spring from an objection to killing
or to war. She did not think that Truman ought not to be hon-
oured simply on the grounds that he was responsible for killing
a lot of people. As we saw in the previous chapter, Anscombe
thought pacifism to be not merely a false, but a positively harmful,
doctrine (Anscombe, 1939, 1957c, 1961). Rather, her objection
was grounded in her conviction that there is a fundamental and
inviolable difference – a difference she viewed pacifism as under-
mining or occluding – between an act of killing and an act of
murder. She thought while it would be sentimental and impracti-
cable to oppose killing, opposition to murder was neither of these
things. To commit murder is to choose to kill the innocent as a
means to an end, or as an end in itself. Truman chose precisely
this, his end being Japan’s unconditional surrender. Like Hitler,
Nero and Genghis Khan he was, thereby guilty of murder; and
so to bestow the degree on Truman, Anscombe argues, would be
‘to pretend … that a couple of massacres to a man’s credit are not
exactly a reason for not showing him honour’ (Anscombe, 1957c,
pp. 64–65).

Let us focus on the idea that what Truman did, in ordering the
atomic bomb to be dropped, was murder and as such was wrong. Is
this true? There are three stages at which one might resist this claim.

First, without relying on the difference between killing and
murder, one might wonder whether it is really correct to say
that Truman killed the Japanese people. Unlike Genghis Khan,
Truman did not, after all, go into battle wielding a weapon. All he
did was give the order. It’s not as if he himself dropped the bomb.
Indeed, we might reflect, there were lots of people in a long chain
of command between his giving the order and thousands of peo-
ple getting burned alive. If we add in all the people who designed
and built the bomb, we can see that Truman was just a tiny cog in
a long causal chain, most of which was utterly outside the reach
of his agency. Surely, we might worry, Truman cannot be fully
responsible for what happened, when we consider that his order
had to pass through all those other agents – any of whom could
have prevented the outcome – before it could result in a button
being pressed in the cockpit of The Enola Gay days later and thou-
sands of miles away.

THREE TASKS FOR INTENTION 31

Anscombe associates this line of thought with a ‘gentle view of
responsibility’, on which ‘to hold someone responsible for what he
did is to ascribe the … causality of it as an event to him’ (1957b,
p. 267). On such a view, to hold Truman responsible for the
deaths of 400,000 people is deeply unfair.

Second, one might also question whether, even if Truman’s act
was killing, it was murder. The definition of murder includes the
condition that the deaths of the innocent must be intended. How
are we to determine whether it was Truman’s intention to kill the
innocent? Usually the idea of intention is connected to desire, and
doing what one intends with enjoyment or pleasure. These psycho-
logical states and emotions seem quite out of place in a description
of Truman. It seems highly likely that he had no desire that thou-
sands should die – indeed, we might add, his order to drop the
bomb was given precisely because he desired to limit the number of
people who would die. Though Truman may well have been happy
that the surrender came, it would be surprising to find he felt pleas-
ure upon learning that, as per his order, thousands were dead.

Finally, even if we accept that Truman was guilty of the sort
of killing that is murder, one might object to Anscombe’s claim
that murder is always prohibited. One might hold that sometimes
killing the innocent is the appropriate means to take to one’s end,
especially if achieving that end ‘pretty certainly save[s] a huge
number of lives’. To think otherwise is a kind of utopianism, dan-
gerous and foolish in the context of the real complexities of war.

In setting out Anscombe’s views about Truman we have already had
to make use of a number of the key concepts in Intention. We have
spoken about: intentions; means and ends; intended, foreseen conse-
quences; actions; intentional actions; desires; and causes. The dispute
we have sketched gives us a hint as to why Anscombe thought that
to defend the view that what Truman did was murder and as such
was wrong, she would need to say something about these concepts.

MODERN MORAL PHILOSOPHY

To bring the task of Intention into clearer focus, we should turn to
two papers written by Anscombe shortly after her failed attempt

THREE TASKS FOR INTENTION32

to block Truman’s degree, and during the time at which she was
writing Intention. The first is a talk given by Anscombe on the
BBC in February 1957, ‘Does Oxford moral philosophy corrupt
the youth?’.2 The second is her famous paper ‘Modern Moral
Philosophy’, published in 1958. In these papers Anscombe con-
nects the view she attacked in her ‘Mr Truman’s Degree’ (1957c)
with what she identifies as the predominant view in moral philos-
ophy of her day.

Posing the question, ‘Does Oxford moral philosophy corrupt
the youth?’, Anscombe answered ‘No’: ‘Oxford moral philoso-
phy is perfectly in tune with the highest and best ideals of the
country at large’ (1957b, p. 164). The tone is ironic. Anscombe
did not intend to praise Oxford moral philosophy, but rather
to take a swipe at the ‘highest and best ideals of the country at
large’. She meant: students of philosophy at Oxford are utterly
corrupted by society at large, and as such there is no further
work of corrupting them for Oxford moral philosophy to do.
For Anscombe, then, the position she took against the prevail-
ing public opinion about Truman – that what he did was rather
heroic – was also a challenge to a prevailing moral philosophy, to
which we now turn.

In ‘Modern Moral Philosophy’ Anscombe presents three
inflammatory theses: that ‘it is not profitable … to do moral phi-
losophy … until we have an adequate philosophy of psychology’;
that ‘the concepts of … moral obligation and moral duty … and of
what is morally right and morally wrong, and of the moral sense of
“ought” ought to be jettisoned’; that ‘the differences between the
well-known English writers on moral philosophy from Sidgwick
to the present day are of little importance’ (1958b, p. 169). As I
suggested in the introduction to this chapter, these theses bear
directly on Anscombe’s three tasks in Intention.

In defence of the third thesis, Anscombe coined the label
‘consequentialism’, to characterise what was common between
all those moral philosophers – that which was ‘corrupt’ in their
thinking and the thinking of society at large. This label has per-
sisted, but something of Anscombe’s original meaning has been
lost. In contemporary ethics the term ‘consequentialism’ has come
to be more or less synonymous with the view that the normative

THREE TASKS FOR INTENTION 33

value of an act is determined by its consequences, and by them
alone. However, when Anscombe coined the term she meant it to
have broader application.3

For Anscombe, the view that ‘“the right action” means the one
which produces the best possible consequences’ (1958b, p. 180)
was a symptom of consequentialism, not critierial of it. Anscombian
Consequentialism – as I will call it here to avoid confusion – is the
thesis that ‘there is no kind of act so bad but it might on occasion
be justified by its consequences, or by the likely consequences of
not performing it’ (Geach, HLAE, p. xvii), and it is the denial of
prohibition rather than the focus on consequences per se to which
Anscombe is opposed. The ‘Hebrew-Christian Ethic’ necessitates
a form of moral absolutism; it says that certain kinds of acts are
prohibited ‘in virtue of their description as such-and-such iden-
tifiable kinds of action’, no matter what the cost of inaction, nor
what the benefits of action (Anscombe, 1958b, p. 182). The Ten
Commandments are categorical: thou shalt not murder. No ifs, no
buts. This ethic does indeed say, ‘You may not do evil that good
may come’ (Romans 3:8).

Anscombian Consequentialism is simply the denial of moral
absolutism, and as such is ‘quite incompatible with the Hebrew-
Christian Ethic’:

For it has been characteristic of that ethic to teach that there are certain
things forbidden whatever consequences threaten, such as: choosing
to kill the innocent for any purpose, however good; vicarious punish-
ment; treachery (by which I mean obtaining a man’s confidence in a
grave matter by promises of trustworthy friendship and then betray-
ing him to his enemies); idolatory; sodomy; adultery; making a false
profession of faith.

(Anscombe. 1958b, p. 181)

Outside the Catholic church, few would agree with Anscombe’s
list of prohibited acts. However, the question of whether the loss of
moral absolutism – leaving the content of its prohibitions open –
in favour of an Anscombian Consequentialist ethics is conducive
to human flourishing, can still be asked. An ethics of prohibition
might, for example, hold that murder is prohibited in virtue of the

THREE TASKS FOR INTENTION34

kind of act that it is, while denying that sex outside heterosexual
marriage is prohibited as such.

Indeed, we need not be religious to feel the pull of the idea
that certain acts are forbidden no matter what. In one version of
the so-called trolley problem, we are asked to consider whether a
doctor might legitimately kill a perfectly healthy person in order
to harvest their organs and thereby save ten – or more if you like –
lives. We think not, because to do so would be murder. Most, I
hope, agree that torturing a person in order to get information out
of her is impermissible, even if that information is very important
and could save many lives. Torture is wrong as such. An individual
for whom murder, rape, lying, theft, were always ‘live options’,
available to be considered as courses of action should circum-
stances favour them, would, it seems clear, be a very dangerous
and disturbing individual.4

Anscombe’s second claim, that ‘the concepts of … moral obli-
gation and moral duty … and of what is morally right and morally
wrong, and of the moral sense of “ought” ought to be jettisoned’
has attracted an enormous amount of criticism, much of it mis-
placed. We can see why it has proved inflammatory if we enlarge
the quote. Anscombe continues by explaining why they should
be jettisoned:

[T]hey are survivals, or derivatives from survivals, from an earlier
conception of ethics which no longer general survives, and are only
harmful without it.

(Anscombe, 1958b, p. 169)

The ‘earlier conception’ is the Divine Law conception of
Christianity, that is, and ethics based on the categorical charac-
ter of the Ten Commandments. Anscombe’s claim here is that
any deontological ethical theory that does not contain a ‘divine
legislator’ is false or incoherent.

Deontology is – certainly was at the time Anscombe
wrote ‘Modern Moral Philosophy’ – the main competitor to
Anscombian Consequentialism. It is the view that the rightness
or wrongness of an action is conformity or conflict with a moral
norm. Divine Law theory is just one species of deontological

THREE TASKS FOR INTENTION 35

theory; other examples are Contractualism and Kantianism.
For philosophers defending these theories, arguments against
Consequentialism do not turn on a simple incompatibility with
the Catholic faith. However, what a Divine Law theory has,
which other versions lack, is a divine lawgiver, a Being who is
the source of the norms and of their binding force. The chal-
lenge for any alternative deontological theory is to explain how a
moral norm can be absolutely binding – can be categorical rather
than hypothetical – in the absence of such a Being. In ‘Modern
Moral Philosophy’ Anscombe gives astonishingly brief argu-
ments against philosophers who seek to derive the norms from
conscience, self-legislation, society, contract (Anscombe, 1958b,
p. 186). Every source of norms other than God, she says, is either
nonsense or dangerous.

Anscombe’s short shrift with an entire philosophical tradi-
tion raises heckles, but what has really enraged philosophers
is their suspicion that Anscombe is offering ‘a version of the
Dostoyevskian claim that if God is dead everything is permit-
ted’ (Blackburn, 2005a). If one takes it that a commitment to
moral absolutism just is a commitment to deontological ethics
(of which the Hebrew-Christian ethic is a species) then
Anscombe’s claim that Divine Law theory is the only respecta-
ble deontological theory, is equivalent to the claim: if you don’t
believe in God then there is no such thing as an ethics of pro-
hibition. Which, by Anscombe’s lights, is to say there’s no such
thing as morality. Blackburn complained that, since Anscombe
did believe in God, she ‘of course, had no intention of jetti-
soning the concepts of moral obligation and duty, which are
needed to frame her other principal claim, which is that certain
things are forbidden, whatever the consequences’ (Blackburn,
2005a). So, morality is good enough for Anscombe and her
fellow Catholics, but not for the rest of us.

This reading of Anscombe’s second thesis is, however, quite
wrong; Anscombe thought that she could frame an ethics of
prohibition without speaking of a distinctive ‘moral sense of
“ought”’ by using the ethical concepts found in Aristotle. This
would generate a version of moral absolutism which was not
deontological in character. Blackburn’s verdict that Anscombe is

THREE TASKS FOR INTENTION36

‘simply wrong’ (Blackburn, 2005a) turns on his mistaken assump-
tion that, for Anscombe, one cannot be a moral absolutist without
being a deontologist. Rather, Anscombe held that if one had an
‘adequate philosophy of psychology’ then a third sort of ethical
theory – a competitor to both Anscombian Consequentialism and
Deontology – emerges, one that is grounded in Aristotle’s account
of the virtues. Mary Geach, responding to Blackburn’s attack
explained as follows:

Anscombe maintains that the class of actions which are illicit
(i.e., contrary to divine law) is the same class as the class of actions
which are contrary to the virtues which one has to have in order to
be a good human being. She did not think one needed a divine law
conception of ethics to know what a good human being was, or what
virtues he had. Aristotle did not speak of divine law, and she saw in
him a figure to whom atheists (as well as Christians) could look as an
example of how to think about vice and virtue.

[…] She wanted people who did not believe in God to stop asking
questions like ‘Is this morally right?’, and to start asking questions
like ‘Is this gluttonous?’ or ‘Is this that kind of injustice which is called
murder?’. She did not think that an atheist could have no desire to
be a good man, or to act well, or that in him such a desire must be
meaningless. She was not attacking atheism as leading to libertinism.

(Geach, 2005)

The claim that ‘the class of actions which are illicit … is the same
class as the class of actions which are contrary to the virtues’ is
an exciting one, and holds out the promise of a secular abso-
lutist ethics which marches in step with one which is Christian.
Ultimately, the question of what makes gluttony vicious and
humility virtuous will be grounded in questions about the
value and purpose of a human life, the answer to which will,
for Anscombe, be located in her faith. The role of God in
Anscombe’s ethics is thus as telos rather than logos. However,
the possibility of a naturalist answer to such a question should
not be ruled out; ‘Modern Moral Philosophy’ precipitated the
revival of virtue ethics and its sibling moral psychology, and the
search for such an answer.5

THREE TASKS FOR INTENTION 37

INTENDED AND FORESEEN CONSEQUENCES

Anscombe says that ‘the distinction between the intended and the
merely foreseen consequences of a voluntary action is … absolutely
essential’ to an ethical theory which ‘forbids a number of things as
bad in themselves’:

if I am answerable for the foreseen consequences of an action or
refusal, as much as for the action itself, then these prohibitions will
break down. If someone innocent will die unless I do a wicked thing,
then on this view I am his murderer in refusing: so all that is left is to
weigh up the options.

(Anscombe, 1961, p. 58)

We can illustrate the distinction that Anscombe wants to preserve
by thinking again about Truman. Suppose Truman is deliberat-
ing about whether or not to give the order to drop the bomb.
Thinking about issuing the order, he notes that if he gives it, the
result will be the deaths of thousands of innocent people. On the
other hand, if he instead refuses to give the order, he foresees that
there will be an ‘Okinawa from one end of Japan to the other’
during which many innocents – Japanese civilians – along with
many US soldiers will die. A consequence of his refusal will thus
be the deaths of thousands of innocent people. He thus reasons:

[A] If I give the order then thousands of innocent people will
die.

[B] If I do not give the order then thousands of innocent people
will die.

[C] Therefore, to decide what to do I should work out which
option will lead to a better situation.

In both [A] and [B] the same description – thousands of innocent
people will die – appears as the consequent of the conditional. Set
out as they are above, there does not seem to be any difference
between these consequents that could be morally relevant, aside
from which set of deaths will constitute, or be part of, a better
overall situation. How many thousands of people will clearly

THREE TASKS FOR INTENTION38

be relevant here, as, perhaps how innocent those people are
(i.e. perhaps [A] includes more Japanese civilians but [B] more
prisoners of war). Questions about the Allies’ war-aims, the future
of US–Japanese relations, the stability of a post-nuclear peace, and
so forth, might also need to be weighed.

If we look carefully, however, we can see that there is another
dimension of difference, one which does not have to do with
weighing the consequences. This difference lies in the way in which
Truman relates to the described event – thousands of innocent
people dying – in [A] and in [B]. We can illuminate this difference
by imagining Truman’s utterance ‘Thousands of innocent people
are going to die’. In [A] this utterance would be an expression of
his intention; in [B] it would be a prediction founded on evidence.
Truman’s reason for thinking that thousands of innocent people
will die if he refuses to give the order – case [B] – is, let us say, that
this was the case at the battle of Okinawa and intelligence suggests
that the Japanese people have not changed their resolve and are
still willing to commit suicide rather than surrender. Based on this
evidence he can predict, perhaps with a good degree of certainty,
how the future will go. In [A], however, that thousands of innocent
people are going to die is not a prediction founded on evidence,
but rather gives Truman’s intention. It is a description of how he
thinks the future should be, whereas in [B] it is his description of
how he thinks the future will be.6 This distinction is, of course,
Anscombe’s topic in the opening and closing sections of Intention.

One of the aims of Intention is to secure the difference between
an expression of intention for the future and a prediction in a way that
shows how different stances we can take toward possible futures
can be relevant to questions about our responsibility for what hap-
pens. If he had thought about the difference between [A] and [B]
in this way, Truman might, for example, have concluded instead:

[C] Given [A] I must not give the order, because intentionally
killing the innocent is murder. So I should either adopt a
different end or consider alternative means. Given [B] it
would be good to do whatever I can to ensure the death-
toll amongst the innocents is as low as possible (e.g. avoid
land-invasion if possible; use all available diplomatic

THREE TASKS FOR INTENTION 39

channels; re-examine the necessity of continuing the war
with Japan; send plenty of aid and medicine; explore pos-
sibilities for communicating assurances to the Japanese
people; reconsider the end of unconditional surrender).

INTENTION WITH WHICH

The legitimacy of the above reasoning depends, as Anscombe says,
on a distinction between the foreseen and intended consequences
of an action. If Truman’s foreseeing the deaths of thousands of
innocents is the same as his intending the deaths of thousands
of innocents, then the distinction between an expression of inten-
tion and a prediction breaks down and Truman may as well go
back to weighing the consequences. This distinction is, Anscombe
notes in several papers (e.g. 1961, 1982b, 1982c, 1983b), at the
core of the Catholic doctrine of double effect. The doctrine says:
if intentionally B-ing is prohibited, and B-ing is a foreseen conse-
quence but not an intended consequence of A-ing, then A-ing is
not necessarily prohibited. So, intentionally killing the innocent,
say, is prohibited. But if killing the innocent is an unintended but
foreseen consequence of my ordering a land invasion, doing so
may nevertheless be permitted.7

The doctrine of double effect requires a distinction between
intended and foreseen consequences. However, even if the distinc-
tion is secured, the doctrine can come in for ‘abuse’ (Anscombe,
1961, p. 58) – abuse which makes it look like as if the following
reasoning might have been available to Truman:

[A’] Giving the order with the intention to kill the innocents is
prohibited.

[B’] But giving the order with the intention to destroy some
buildings in Hiroshima is not prohibited, even if I foresee
the deaths of the innocent.

[C’] Therefore I will give the order with the intention to destroy
some buildings.

Anscombe calls this kind of reasoning ‘double-think about double
effect’ (Anscombe, 1961, p. 58).

THREE TASKS FOR INTENTION40

In Intention she argues that the idea that ‘one can determine
one’s intentions by making such a little speech to oneself if obvi-
ous bosh’ (§25, p. 42). The difficulty we have in recognising this
‘bosh’ when we see it is, Anscombe argues, connected with a par-
ticular dominant conception of intention.

From the seventeenth century till now what may be called Cartesian
psychology has dominated the thought of philosophers and theolo-
gians. According to this psychology, an intention was an interior act
of the mind which could be produced at will. Now if intention is all
important – as it is – in determining the goodness or badness of an
action, then on this theory of what intention is, a marvellous way
offered itself of making any action lawful. You only had to ‘direct your
intention’ in a suitable way. In practice, this means making a little
speech to yourself: ‘What I mean to be doing is …’.

(Anscombe, 1961, p. 59)

This account of intention makes a mockery of the doctrine of
double effect by making it always possible to avoid prohibition. A
second aim in Intention is to undermine the idea that the intention
with which you act (Anscombe’s third topic in Intention) can be
determined by a private ceremony in which one says to oneself or
out-loud ‘I am giving this order in order to destroy some buildings
(and not in order to kill these people)’. Anscombe wishes to say
instead: ‘Roughly speaking, a man intends to do what he does’
(§25, p. 45).

AN ACCOUNT OF ACTION

Anscombe, we have seen, thought that ‘the class of actions which
are illicit (i.e., contrary to divine law) is the same class as the class
of actions which are contrary to the virtues which one has to have
in order to be a good human being’. I have suggested that we see
Anscombe’s first thesis (in ‘Modern Moral Philosophy’) as relat-
ing to this idea. Anscombe thought that what ethics needed was
an account of the class of actions which are contrary to the vir-
tues, and that such an account requires ‘an adequate philosophy
of psychology’. Anscombe describes the project:

THREE TASKS FOR INTENTION 41

[A]n explanation is required how an unjust man is a bad man, or an
unjust action a bad one; to give such an explanation belongs to ethics;
but it cannot even be begun until we are equipped with a sound phi-
losophy of psychology. For the proof that an unjust man is a bad man
would require a positive account of justice as a ‘virtue’. This part of the
subject-matter of ethics is, however, completely closed off to us until
we have an account of what type of characteristic a virtue is – a problem
not of ethics but of conceptual analysis – and how it relates to the
actions in which it is instanced … For this we certainly need an account
at least of what a human action is at all, and how its description as ‘doing
such and such’ is affected by its motive and by the intention and intentions
in it; and for this an account of such concepts is required.

(1957b, p. 174, emphasis added)

This description of the project is dense and difficult, but we can
begin to see what sort of account Anscombe might be talking
about if we look to her discussion of the second thesis in ‘Modern
Moral Philosophy’.

So far we have considered the role of a ‘Cartesian psychology’,
according to which ‘an intention was an interior act of the mind
which could be produced at will’. In the first thesis, the target
is not Cartesian but Humean psychology. Hume drew to phil-
osophical attention the difficulties that attend a certain kind
of reasoning which seems central to ethical argument. The rea-
soning is that which moves from statements about what is the
case – descriptive statements which might be discovered and
justified by observation – and statements about what ought to
be the case – prescriptive or normative statements whose discov-
ery seems to lie outside the realm of science. Hume complained
that moral philosophers tended to move from the former to
the latter without offering any account of ‘what seems alto-
gether inconceivable’; namely, how ‘this new relation can be
deduced from others that are entirely different from it’ (Hume,
1739/2007, p. 302).

Hume’s reasons for finding this move ‘altogether inconceivable’
are rooted, at least in part, in his account of human psychology,
and the deep separation he made between reason and passion; it is
far outside the scope of this Guidebook to discuss this here. What

THREE TASKS FOR INTENTION42

should be noted, however, is that the Humean idea – that such a
transition was inconceivable – was incredibly influential amongst
Anscombe’s contemporaries. The idea of an ‘is-ought gap’ or a
‘fact-value distinction’, and the associated so-called ‘naturalistic
fallacy’ (Moore, 1903), when viewed through the lens of linguistic
philosophy, became the thesis that ethical statements are non-
factual: when I say ‘Murder is wrong’ or ‘You ought not to do that,
it’s stealing’, I am not saying something that is true or false, I am
not stating a fact, rather I am expressing disapproval or communi-
cating to you my emotional response.

Responding to this, Anscombe asks how someone who
insisted that ethical claims were not the sort of thing that could
be true or false, might make sense of the idea that I owe the
grocer for some potatoes he has delivered me. Such a philos-
opher, she proposes, would find themselves in the position of
having to admit that while it was true that I ordered potatoes,
the grocer supplied them, and he sent me a bill, the statement ‘I
owe the grocer £2’ was not the sort of thing that could be true
or false (1958b, p. 172; 1958c, p. 22). But, says Anscombe,
‘it would be ludicrous to pretend that there can be no such
thing as a transition from, e.g., “is” to “owes”’ (1958b, p. 172);
I ordered potatoes and the grocer delivered them and as such
I owe the grocer for the potatoes. Her thought is that if the
transition from is to owes is legitimate, the transition from is to
ought surely is too.

Anscombe’s point here is not flatfooted. She is not accusing
Hume – or her contemporaries – of missing what is blindingly
obvious. Rather, the case with the grocer is designed to illumi-
nate the complex background conditions which must be in place
before the transition from is to owes is possible. These are not
merely conditions that must exist at the moment at which the
action is performed – for example, the thoughts and intentions of
the protagonists – but include facts about the ‘natural history of
mankind’. As she puts it, the statement that I owe the grocer £2
‘requires [our] or very similar institutions as background in order
so much as to be the kind of statement it is’ (1958c, p. 22).

Anscombe’s thought is that the transition from is to owes, or
from is to ought, is indeed inconceivable if we limit our attention

THREE TASKS FOR INTENTION 43

to a single vignette, in which two agents, located outside of human
history, human practice, human institutions, seek to draw a nor-
mative conclusion from descriptive premises. However, if we place
those agents back in their proper context, our context, we can see
that for them as for us, in the course of normal proceedings it fol-
lows from fact that I ordered the potatoes and the grocer delivered
them that I owe him £2.

Anscombe uses the word ‘brute’ to describe this asymmetric
relation between descriptions of facts: as compared with the fact
that I owe the grocer £2, that he supplied me with potatoes is
brute; similarly, as compared with the fact that he supplied me
with potatoes, that he left potatoes at my house is brute. As we
move between a pair of descriptions, where the first is brute
relative to the second, what we take into account is the context –
which includes the background institutions and practices – and
this context legitimates the move to a higher level of description.
We can add: that the grocer supplied me with potatoes is brute as
compared with the fact that I should pay him.

The relation of ‘brute relative to’ is, says Anscombe, a complex
one. The difficulty is that the move from lower to higher-level
descriptions is not always legitimate. Not every case of my order-
ing potatoes and the grocer delivering them can be described as his
supplying me with potatoes: for example, the order and delivery
might have been ‘arranged as part of an amateur film production’
(1958c, p. 23). Anscombe says that ‘there can be no such thing
as an exhaustive description of all the circumstances which could
theoretically impair the description of an action of leaving a quar-
ter of potatoes at my house as “supplying me with a quarter of
potatoes”’; rather, the relation ‘presupposes a context of normal
procedure’ (1958c, p. 23).

As we will see, the idea that a description of a set of facts can
be ‘brute’ relative to another description of those facts, is central
to Anscombe’s account of action, in which she describes a ‘special
kind of multiplicity of levels of human acts’:

I put ink on paper in the form of letters. I am writing something. I
am in fact signing something with my name. And I’m thereby joining
in a petition to the governor of the state – or prison – where I am

THREE TASKS FOR INTENTION44

important. I am taking part in a campaign to get people tortured
under interrogation. In doing this I am keeping a promise. I am
avoiding trouble with some conspirators who have got me to prom-
ise to do that.

(Anscombe, 1993, p. 149)

We can imagine a similar list of descriptions relating to Truman.
He put ink on paper in the form of letters. He is writing some-
thing. He is in fact signing something with his name. He is
thereby giving the order to bomb Hiroshima. We could say, in
the context of our institutions what it is to give the order to bomb
Hiroshima is to sign this piece of paper in such-and-such circum-
stances (circumstances which will include facts about institutions
like the government and the military, and about war, political
office, nation states, atom bombs, and so forth).

Returning to our question about good or virtuous action, we can
see how getting descriptions like ‘X owes Y money’ or ‘X is sign-
ing a petition to get someone tortured’ or ‘X is signing an order
to get some innocent people killed’, to be descriptions of what is
happening – descriptions of the facts – can be a step toward giving
us what Anscombe says we need:

That I owe the grocer such-and-such a sum would be one of a set
of facts which would be ‘brute’ in relation to the description ‘I am a
bilker’. ‘Bilking’ is of course a species of ‘dishonesty’ or ‘injustice’.

(1957b, p. 174)

If we can understand how it is that what is happening is describ-
able as ‘doing such and such’ we will get to see how it is that an
action is describable, qua ‘doing such and such’ as good or bad.
Hume says: you can’t get an ought from an is; the deontologist,
accepting this, looks for a source of the ‘ought’; Anscombe says:
if you describe the facts using certain concepts then the ‘ought’
(or ‘ought not’) is, as it were, already in them.

It will get clearer as we work our way through Anscombe’s
account, given in Intention, of ‘what a human action is at all, and
how its description as “doing such and such” is affected by its

THREE TASKS FOR INTENTION 45

motive and by the intention and intentions in it’. This, of course,
is the second of Anscombe’s topics: intentional action.

We have identified the three tasks for the philosophy of psy-
chology that Anscombe thought we needed to undertake before
it would be profitable to do moral philosophy. They were: an
account of the distinction between an action’s intended and its
merely foreseen consequences; a criterion for the intention with
which a person acted which showed that and why a mere ‘act of
will’ on the part of the agent is not sufficient to determine the
intention with which an action was done; an account of human
action – voluntary and intentional – which could make sense of
the idea that actions can instantiate virtues (that is, that a particu-
lar act, say, my handing you a piece of paper, can be an action of
kindness, or justice, or honesty). One way to fulfil these tasks would
be to study a man, another to study man in his circumstances. I
suggest that Anscombe’s project is the latter. Anscombe says that
‘It belongs to the natural history of man that he has a moral envi-
ronment’ (Anscombe, undated b, p. 224). If this is right – if the
moral environment is, as Murdoch put it, ‘the reality that sur-
rounds man’ then a description of that ‘reality’ would seem to be
a good place to start.

NOTES

1 Readers unfamiliar with contemporary moral philosophy may find ref-
erence to boiling a baby rather grotesque. In fact, unthinkably horrifying
scenarios are part of the landscape of moral thinking for philosophers.
Anscombe suggests that such examples are Symptomatic of a corrupting
philosophy.

A … point of method I would recommend to the corrupter would
be this: concentrate on examples which are either banal: you have
promised to return a book, but … and so on; or fantastic: what you
ought to do if you had to move forward and stepping with your right
foot meant killing twenty-five fine young men while stepping with
your left foot would kill fifty drooling old ones.

(1957b, p. 163)

THREE TASKS FOR INTENTION46

2 Given the continued ubiquity of Oxford politics, philosophy and eco-
nomic graduates amongst the ruling elite, now might be a good time to
revisit this question.

3 See Cora Diamond (1997) for a helpful discussion.
4 Jack Bauer, in the American television series 24 is perhaps one such

individual.
5 For outstanding examples, see Philippa Foot’s Natural Goodness (2001)

and Michael Thompson’s (2008) Life and Action: Elementary Structures of
Practice and Thought.

6 See Vogler (2012) for an excellent discussion of these two ways of relat-
ing to the future and their relevance to anti-Consequentialism.

7 Note that this is a negative claim. The doctrine of double effect does not
say what one can or should do, but only that the act in question is not
prohibited on the grounds of the kind of action that it is; this is not to say
that it may not be prohibited on other grounds nor that another kind of
action might not be better.

SUGGESTED READING

Anscombe, G. E. M. (1957b). ‘Does Oxford Moral Philosophy Corrupt the Youth?’.
Radio broadcast, printed in The Listener, Vol. 57 (14 February), pp. 226–227.
Reprinted in HLAE, pp. 161–168.

—— (1957c). ‘Mr Truman’s Degree’. Pamphlet published by author. Reprinted in
ERP, pp. 62–71.

—— (1958b). ‘Modern Moral Philosophy’. Philosophy, 53, pp. 1–19. Reprinted in
HLAE, pp. 169–194.

Diamond, Cora (1997). ‘Consequentialism in Modern Moral Philosophy and in
“Modern Moral Philosophy”’. In David S. Oderberg and Jacqueline A. Laing (eds),
Human Lives: Critical Essays on Consequentialist Bioethics (Palgrave Macmillan),
pp. 13–38.

Foot, Philippa (2001). Natural Goodness (Oxford: Oxford University Press).
MacIntyre, Alisdair (1990). Three Rival Versions of Moral Enquiry: Encyclopaedia,

Genealogy, and Tradition (London: Duckworth).
Richter, Duncan (2011). Anscombe’s Moral Philosophy (Plymouth: Lexington).
Teichmann, Roger (2008). The Philosophy of Elizabeth Anscombe (Oxford: Oxford

University Press).
Vogler, Candace (2012). ‘In Support of Moral Absolutes’, Villanova Law Review, 57(5),

pp. 893–906.

3
‘THE SUBJECT UNDER

THREE HEADS’

We have looked at the genesis of Intention and were able to identify
the sort of thinking about intention and intentional action that
Anscombe wanted to undermine. One target we introduced was
‘Cartesian psychology’ according to which ‘an intention was
an interior act of the mind which could be produced at will’.
Anscombe, we saw, thought that this distinctively Modern doc-
trine left the doctrine of double effect open to abuse, and made
it difficult – perhaps impossible – to see how an action rather
than an intention could be good or bad, virtuous or vicious. It
is natural, says Anscombe, to think that questions about what a
man intends, and what his intentional actions are, are ultimately
settled not by what he does, but by what is going on in his mind.1
However natural it may be, Intention aims to show that it is false.

Anscombe’s hostility to this psychology has been generally
recognised, and this is one thing that has made her book so
attractive to philosophers of mind in recent years. Those in the
contemporary neo-Kantian tradition who, following Strawson

‘THE SUBJECT UNDER THREE HEADS’48

(1959, 1966), want to insist that it is persons and not minds that
are the subjects of psychological predicates, have been especially
drawn to the central role that Anscombe gives in her book to
actions and bodily movements (e.g. McDowell, 2011). Anscombe
says that it is a mistake to ‘think that if we want to know a man’s
intentions it is into the contents of his mind and only into
these that we should enquire’ and proposes instead that the first
thing to consider is ‘what a man actually does’ (§4, p. 5). This
is reflected in the emphasis that Anscombe gives to intentional
action over intention for the future in her book. Anscombe ded-
icates 19 of the 52 paragraphs of her book (§§4–21) explicitly
to the topic of intentional action, and another 27 (§§22–49)
to the intention with which an action is done. Expressions of
intention for the future warrant only five sections of discussion
(§§2–4 and 50–52).

In this chapter I want to say something about the kind of prior-
ity that Anscombe gives intentional action in her monograph, and
through doing so to provide a framework for the book.

THE SUBJECT INTRODUCED UNDER
THREE HEADS

Anscombe’s entry in her contents for §1 is: ‘The subject intro-
duced under three heads: expressions of intention for the future,
intentional action, and intention in acting’ (contents, §1, p. i).
She begins her book:

Very often, when a man says, ‘I am going to do such and such’, we
should say that this was an expression of intention. We also some-
times speak of an action as intentional, and we may also ask with what
intention the thing was done.

(§1, p. 1)

She continues:

[N]ow if we set out to describe this concept, and took only one of these
three kinds of statement as containing our whole topic, we might very
likely say things about what ‘intention’ means which it would be false

‘THE SUBJECT UNDER THREE HEADS’ 49

to say in one of the other cases …. Realising this might lead us to
say that there are various senses of ‘intention’, and perhaps that it
is thoroughly misleading that the word ‘intentional’ should be con-
nected with the word ‘intention’ … Where we are tempted to speak of
‘different senses’ of a word which is clearly not equivocal, we may infer
that we are in fact pretty much in the dark about the character of the
concept it represents.

(§1, p. 1)

The ‘three heads’ identified here then divide the book into three
parts. The first head, expressions of intention for the future, is
the topic of §§2–3 and §§50–52; the second head, intentional
actions, is the topic of §§4–21; the third head, intention with
which, is the topic of §§22–49.

It is clear from what Anscombe says in §1 that she means to
reject any account of intention that does not present these three
heads of intention as aspects of a single concept. However, there
is a question to be asked about what would constitute a unified
account. How we answer this question will profoundly affect the
way we think about the priority that Anscombe gives to ‘what a
man actually does’ (§4, p. 5) in her book.

THE ‘CONNECTIVE’ APPROACH2

One way to understand §1 is to see it as asking: what is the con-
nection between my intention to act, my intentional action, and
the reason which justifies or explains that action? I will call this
‘Anscombe’s Question’ – though, to be clear, we will see that
Anscombe herself neither poses nor answers this question. As I
explained in Chapter 1, this question is one of three that dom-
inate post-Intention work in the philosophy of action. A unified
account of the concept of intention, according to those who seek
an answer to Anscombe’s Question, would be one that described
the connections between the phenomena that are picked out
by the concept in each of its three guises. In doing so it would
explain how, for example, my having the intention to go to the
library is connected – causally, explanatorily, rationally – to my
going to the library.

‘THE SUBJECT UNDER THREE HEADS’50

An intuitive account of the phenomena to be connected in an
answer to Anscombe’s Question can be given by appeal to a few
examples. An intention for the future is a current mental state. I can
now have the intention to finish this chapter by Tuesday, to call
my mother at the weekend, to own a Ferrari by the time I am 65.
Intentional actions are a species of action, and actions are a species
of event. My arm goes up. If it goes up under my volition and not,
say, via a mechanism or in my sleep, its going up is an action of
mine: I raised my arm. Sometimes, when it is an action, it is an
intentional action. With arm-raisings the distinction between action
and intentional action is not easy to make out – it is hard to describe
a case in which one would say that I raised my arm but did not
do so intentionally – but this distinction comes clearly into view
if we fill in a little more context. (Later, we will be able to see just
why the distinction doesn’t get easy traction when we are describing
bodily movements.) Suppose the arm-raising to be a case of voting
in favour of repealing the fox-hunting ban; you might ask: ‘Did you
intend to vote in favour when you raised your arm?’. A negative
answer need not imply that my arm-raising had not been an action
of mine. It might be that I raised my arm intending to vote against
the motion, having misheard the chair’s instruction. A full account
of intentional actions would need to describe the properties that dis-
tinguish intentional actions from other kinds of action – involuntary,
accidental, unintentional, and so forth. The intention with which I
act is my reason for acting. It is that which figures in my practical
reasoning about what to do, and that which occurs in rationalising
explanations. For example, ‘she raised her arm because she wanted to
vote in support of the motion’, or ‘she is writing a book on Intention
because she wants a Ferrari and believes that by writing a philosophy
book she can acquire the relevant riches’.

Many philosophers of action have sought, are seeking, to
answer Anscombe’s Question. Indeed, answering Anscombe’s
Question has some claim to be ‘the principal task of the philos-
ophy of intention’ (Setiya, 2014). This, as we saw, is one way in
which Intention has come to be an important book: Anscombe is
credited with introducing the topic of intention under its three
heads, and thereby setting the research agenda for contemporary
philosophy of action.3

‘THE SUBJECT UNDER THREE HEADS’ 51

Those working on answering Anscombe’s Question have found
that a good way to go about answering it is to isolate one of the
cases as explanatorily prior to the others. Once an account has
been given in that case, the others can then be explained by refer-
ence to it. They have found that the natural place to start is with
intention for the future.

Why is this the natural place to start? In a surprisingly influen-
tial argument, Donald Davidson proposed out that any complete
answer to Anscombe’s Question must include a description of the
mental state of ‘pure intending’, where to have a ‘pure intention’
is to have an intention which does not issue in action and which
cannot be explained or rationalised. Davidson gives an example:

Someone may intend to build a squirrel house without having decided
to do it, deliberated about it, formed an intention to do it, or reasoned
about it. And despite his intention, he may never build a squirrel
house, or do anything whatever with the intention of getting a squirrel
house built.

(Davidson, 1978, p. 83)

If having an intention to act always led to action, or always came
about as the result of reasoning, then there would be hope for
giving an account of it in terms of those other phenomena. But
since the potential squirrel-house builder case is intelligible, says
Davidson, we must admit pure intending into our theory. And
once one has done that ‘there is no reason not to allow that inten-
tion of exactly the same kind is also present when the intended
action eventuates’ (Davidson, 1978, p. 89).

It is worth noting that Davidson’s example is extremely strange,
and that it is not clear that such a case is genuinely intelligible. If
I was wondering whether you intended to build a squirrel house
and discovered that you had never decided to build one, never
deliberated about whether to build one, never formed an inten-
tion to build one, and never reasoned about building one; and
if you never did anything toward building one even though, say,
you had all the materials you would need to hand and plenty of
opportunity to commence the project, that would seem to settle
it that you did not intend to build a squirrel house. Your avowal

‘THE SUBJECT UNDER THREE HEADS’52

to the contrary – ‘I intend to build a squirrel house’ – would be
unintelligible except as a joke. As we will see, this kind of unintel-
ligibility is one of the topics of Intention (esp. §18).

I suspect that what makes Davidson’s argument seem persua-
sive is the fact that we already find the idea of ‘pure intending’
extremely compelling; therefore, we don’t need much convincing
that it exists. This is due in part to two quite everyday features of
the concept of intention. The first is that I can have intentions on
which I never act. I can intend to go to the cinema on Tuesday and
be run over by a bus on Monday before I even get so far as looking
up the film listings. The second is that ‘No particular reason’ or
‘I don’t know why’ are sometimes intelligible answers to the ques-
tion ‘Why are you doing that?’ or ‘Why do you want to do that?’
(cf. §17). These features do not themselves imply the existence
of pure intending – as we will see, Anscombe’s gives an account
of them which makes no appeal to any such state. However, for
reasons that Intention explores, we find it intuitive or natural to
explain these cases in terms of such a state. One reason is the
inheritance of Cartesian psychology.

Someone who begins her answer to Anscombe’s Question with
intending to act, clearly endorses the view that ‘if we want to know
a man’s intentions it is into the contents of his mind and only into
these that we should enquire’ (§4, p. 5). Because Anscombe quite
explicitly denies this – along with the Cartesian psychology that
she thinks can underpin it – if we read Intention as an answer to
Anscombe’s Question, we must read it as a rejection of this intui-
tive view. Michael Bratman explains:

Instead of beginning with the state of intending to act [some theo-
rists] turn immediately to intention as it appears in action: [they] turn
directly to acting intentionally and acting with a certain intention …
This is, for example, the strategy followed by Anscombe in Intention.

(Bratman, 1987, p. 5)

I do not want to comment directly on the merits of attempting
to reverse the order of explanation, but rather to focus on what
happens to a reading of Intention if it is understood as engaged
in that project; that is, the project of answering Anscombe’s

‘THE SUBJECT UNDER THREE HEADS’ 53

Question by affording explanatory priority to intentional action.
Anscombe’s book becomes at worst incoherent, at best suggestive
but incomplete.

A first, and glaring, problem for someone who wants to insist
that this is an adequate interpretation of Anscombe’s thought is
that intention for the future is not, in fact, Anscombe’s first head.
The first head is the verbal expression of intention for the future:
‘Very often, when a man says, “I am going to do such and such”,
we should say this was an expression of intention’ (§1, p. 1). Moran
and Stone point out with some understatement: ‘In general, that
“expression of Φ” will be pertinent in studying a psychological
concept Φ isn’t simply to be taken for granted’ (Moran and Stone,
2011, p. 34; see also Teichmann, 2014a). And this is precisely what
Anscombe would have to be taking for granted if she meant her
investigation into expressions of intention to be equivalent to an
investigation into the mental state of intending. The awkwardness
of this reading ramifies through §§2–4 of Intention as Anscombe
puzzles over the distinction between the expression of intention
and prediction; surely the investigation should say something
about intentions themselves – i.e. what expressions of intention are
expressions of – and not focus instead on a difference between two
kinds of statement? Later, it becomes impossible to make sense of
Anscombe’s argument in §20 that ‘the notion of intentional action
would be a very thin one’ if there were no such thing as expression
of intention for the future (§20, p. 32), or her insistence that there
is no natural expression of intention, though we can speak of an
animal’s intentional actions (§2, p. 5; §47, pp. 86–87). So it is not
just that replacing ‘verbal expression of intention for the future’
with ‘intention for the future’ or ‘intending to act’ requires one to
ride roughshod over §1; it also requires one to ignore, discard, or
distort, some of the key discussions later in the text.

Even if one does find a way to justify to oneself the assumption
that every time Anscombe speaks of the verbal expression of inten-
tion for the future what she means to be speaking of is the mental
state of intending to act, this interpretation faces significant prob-
lems. There is not enough discussion of such a state in Intention to
make out how Anscombe proposes to account for ‘pure intending’,
as Davidson calls it. If Anscombe’s book is a contribution to an

‘THE SUBJECT UNDER THREE HEADS’54

anti-Cartesian – or at least, anti-interiorising – project in the
form of an answer to Anscombe’s Question which gives priority
to intentional actions, then she needs either to give an account of
‘pure intending’ or to deny that there is such a thing. Anscombe’s
book says nothing that would fulfil the former role. Moran and
Stone (2011) identify two approaches commentators have taken
to this silence. First, they have berated Anscombe for not giving
a clear account – something that would quite clearly be a major
oversight. Second, they have suggested that Anscombe might be a
behaviourist and hence she might mean to deny that there is such
a thing as ‘pure intending’ (pp. 42–44).

Moran and Stone point out, quite rightly, that a behaviourist
reading of Intention won’t do. Anscombe is quite relaxed about
the existence of unmanifested intentions; she acknowledges what
we have already noted: that ‘a man can form an intention which
he then does nothing to carry out … but the intention itself can
be complete though it remains a purely interior thing’ (§4, p. 9).
She allows mental events (§10), secret thoughts that never mani-
fest (§27, p. 48), cases in which only the agent himself can know
what he intended (§27, p. 48) and says that ‘people [ought to
be] less contemptuous of phenomenalism that it [is] fashionable
to be’ (§28, p. 49, fn 1). She confirms what we know: that I can
intend to do something – just as I can do something – for no
particular reason (esp. §51, p. 90). To fill the gaps in Intention
philosophers must help Anscombe out – perhaps she didn’t see
the problem? – and come up with their own accounts to complete
her unfinished offering.

The next problem for this reading is to locate an account of
intentional action in Intention that explains when an event is an
action, and when an action intentional. There are lots of points
at which Anscombe comes close to doing so, but when one
tries to pin the explanation down, they all seem to disintegrate.
Anscombe somehow starts too far downstream; she doesn’t justify
crucial premises, fails properly to clarify central concepts, and is
skimpy on arguments where they are needed most. Anyone who
has attempted to read Intention through this lens will be familiar
with the sense of frustration – bordering on fury – of Anscombe’s
seemingly slapdash approach to explaining herself or justifying

‘THE SUBJECT UNDER THREE HEADS’ 55

her most important claims, and by her tendency to leave off a
topic at a crucial moment or to lead one though an argumentative
maze which terminates in a brick wall.

Finally, there is the problem of making sense of the book as
a whole. We have already mentioned the oddity of Anscombe’s
decision to introduce her first head by comparing expressions of
intention for the future to predictions (and, indeed, orders), and
the frustration of Anscombe’s seemingly meandering reflections.
It is also the case that some fairly important-looking paragraphs
are quite unamenable to this interpretative frame. For example,
between §47 and §50 Anscombe describes a ‘form of description
of intentional actions’ and descriptions ‘that are dependent for
their existence of this form for their own sense’ (§47, p. 85). Or
Anscombe’s final remark about St Peter: that his case shows that
a certain kind of ‘ignorance’ can make it possible to ‘be as cer-
tain as possible that one will do something, and yet intend not
to do it’ (§52, p. 94). These are just a few points of frustration;
readers will no doubt be able to highlight dozens of other places
where what Anscombe says, or the direction in which she takes
her argument, is utterly baffling and contrary to expectation. Why
does Anscombe write her book in such a strange style, full of non-
sequiturs, asides, and rambling interludes? Why use the conceit of
the question ‘Why?’ and not just set out her position? One will
find oneself asking over and over ‘Why is she talking about that?’,
‘Why is this important here?’, ‘Sure, but so what?’.

Two things have tended to happen in the face of these, and
other, unanswerable interpretative questions (that is, unanswerable
from within this framing). First, philosophers interested in reject-
ing the natural answer to Anscombe’s Question have been drawn
toward philosophers who, inspired by Intention, have sought to
offer such an account, freed from – what Michael Thompson has
dismissed as – the ‘jargon’ and ‘peculiar obsessive theoretic tics’
that characterise Anscombe’s exposition (2011, p. 198). For many
years Donald Davidson’s work was the first place to look look
(in particular, his 1963, 1971 and 1978); in years to come I expect
it will be Thompson’s. Anscombe’s text thus becomes of interest
to historians of philosophy only. Second, philosophers have grav-
itated, en masse, toward a small number of paragraphs, ideas and

‘THE SUBJECT UNDER THREE HEADS’56

slogans, which are then extracted from the text and either debated
at length or pulled out of hats like rabbits, to do work in other
forums. The favoured sections have tended to be between §4 and
§32, and especially those relating to the knowledge that a man has
of his intentional actions, which Anscombe says is ‘without obser-
vation’ (§8, p. 14) and ‘practical’ (§32, p. 57).

This cherry-picking would be understandable, indeed neces-
sary, were Intention a disorganised series of abortive investigations,
incomplete arguments, and inconclusive lines of thought which,
though rich with suggestive passages and flashes of insight, require
first systematising and then completing. However, it is not.

‘WE ARE IN THE DARK ABOUT THE CHARACTER
OF THE CONCEPT’ (§1)

No interpreter can hope to make Anscombe’s book straightforward
or easy to read. But I do want to show that an alternative which at
least renders it coherent and – by its own lights at least – complete
comes to be available once we take care to think about Anscombe’s
philosophical method. When we do this, it becomes apparent that
Intention is not an answer to Anscombe’s Question any more than
the final paragraphs of part I of Philosophical Investigations are an
answer to ‘Wittgenstein’s Question’ (see note 3 above). Rather,
Anscombe – like Wittgenstein – must be understood as seeking
to uncover and reject the framework of philosophical assumptions
that give rise to such a question.

Above I said that it was clear from §1 that Anscombe means
to reject any account of intention that does not present the uses
of the word ‘intention’ as part of a single concept. I said there
was a question to be asked about what would constitute a uni-
fied account of the concept, and I want now to show why an
account which explains the connections between a mental state,
a physical event, and a reason – an account, that is, that answers
Anscombe’s Question – would not count as ‘unified’ by her lights.
As far as Anscombe is concerned, any philosopher who says ‘there
are intentions, and intentional actions and reasons for actions, but
we don’t know how these things are related ’ is someone who is
treating the word ‘intention’ as if it were equivocal in meaning.

‘THE SUBJECT UNDER THREE HEADS’ 57

He is treating it as equivocal because he is acting as he can identify
phenomena associated with the word in the three cases and then
ask: how do these phenomena connect? The truly radical proposal
of Intention is: if you understood the character of the concept to
which ‘intention’ referred, answering Anscombe’s Question would
no longer look like a coherent project.

The idea that we don’t understand the character of the concept
to which ‘intention’ refers is at the heart of Intention, but the point
is a difficult one to grasp. There are two related claims, neither
of which is Anscombe’s. The first is that when we use the word
‘intention’, in our day-to-day activities, we are, or might be, using
it wrongly. We might be applying the concept where it does not
apply, and withholding it where it should be applied. Anscombe
quite clearly does not think this. She insists that the distinction,
e.g., between an expression of intention and a prediction ‘is intu-
itively clear’ (§2, p. 1). We can, she says, usually say ‘straight off ’
and ‘with a great deal of certainty’ what a man’s intentional actions
and intentions are (§4, p. 8). We can recognise ‘examples of the
involuntary’ (§7, p. 8) and we know that when a person jumps at
the sound of a balloon bursting, the noise caused him to jump and
did not give him a reason to jump. Indeed, Anscombe’s method
in Intention takes as its starting point our capacity to employ the
concept of intention and to recognise, at least in the context of
normal proceedings, what a man’s intentions and intentional
actions are. In §2, Anscombe says:

the distinction between an expression of intention and a prediction is
generally appealed to as something intuitively obvious. The distinc-
tion is intuitively clear, in the following sense: if I say ‘I am going to
fail in this exam.’ and someone says ‘Surely you aren’t as bad at the
subject as that’, I may make my meaning clear by explaining that I was
expressing an intention, not giving an estimate of my chances.

(§2, pp. 1–2)

The second thing one might mean if one said ‘we don’t under-
stand the character of the concept to which “intention” refers’ is
that we do not have an understanding of the nature of intentional
actions or intentions. This second claim is one that a philosopher

‘THE SUBJECT UNDER THREE HEADS’58

engaged in the connective project takes as his starting point. His
thought is: we can use our intuitive understanding of the word
‘intention’ to ‘roughly outline’ the relevant phenomena, but hav-
ing done so we need to give an account of their nature. The task is
to give an account of each class, now distinguished, and the place
to look is at members of the class. What properties do members
of the class have in common? This investigation, according to the
connective project, is the work of the philosopher.

To understand the sense in which Anscombe herself thinks we are
‘in the dark about the character of the concept’ we need to find
a way of squaring her remark that ‘[t]he distinction is intuitively
clear’ with the following comment, from the next paragraph:

[Dead-ends] are reached in consequence of leaving the distinction
between estimation of the future and expression of intention as some-
thing that just is intuitively obvious.

(§3, p. 6)

What would it be to acknowledge that the distinction is intu-
itively obvious and yet deny that the distinction can be left as
something that is intuitively obvious?

The idea that we can be in the dark about how our language
operates while at the same time being perfectly at home oper-
ating with it, will be familiar to readers of Wittgenstein. I want
to suggest that the kind of ‘darkness’ Anscombe is talking about
in §1 is that which we are in before we have undertaken the
task of describing the look of a life that contains the concept
‘intention’. To see how it might be possible to know how to use
the word ‘intention’ but not to be able to describe the look of a
life that contains the concept, we need to say a little about what
Anscombe and Wittgenstein mean when they talk about what it
is to possess a concept.

Speaking of the concept length Anscombe remarks:

The competent use of language is a criterion for the possession of
concepts symbolized in it, and so we are at liberty to say: to have
such-and-such linguistic practices is to have such-and-such concepts.

‘THE SUBJECT UNDER THREE HEADS’ 59

‘Linguistic practice’ does not mean merely the production of words
properly arranged into sentences on occasions which we vaguely call
‘suitable’. It is important that it includes activities other than the pro-
duction of language, into which a use of language is interwoven. For
example, activities of measuring, or weighing, of giving and receiving
and putting into special places, of moving about in a huge variety of
ways, of consulting tables and calendars and signs and acting in ways
connected with consultation. It is plausible to say that we would have
no concept of length apart from some activity of measuring, and no
concept of precise comparative lengths of distant objects if the activity
of measuring had not a quite elaborate use of words interwoven into it.

(Anscombe, 1976a, p. 117)

Suppose someone wanted an account of the difference between
an estimate and a measurement of length. One way to give such
an account would be with an example: ‘That queue is around
15m long’; ‘This stick is precisely 12.4cm’. This wouldn’t be much
good unless the question had been asked by someone who already
had the concepts of an estimate and a measurement, but did not
know the meaning of the words ‘estimate’ and ‘measurement’. For
example, a German speaker who knew schätzen and messen but
not the English equivalents. Another way to answer the question
would be by definition: ‘An estimate is a guess at something’s
size, but a measurement involves using instruments, like rulers
and so forth’. This explanation might do for a child who already
had some of the relevant concepts, already knew about rulers and
some of the activities associated with measuring an estimating;
like buying things by weight in a shop, or timing a race, or making
a mark on a height chart.

But if we ask in philosophy, as she puts it in §2 (p. 2), Anscombe
thinks we need an account which includes what we take for granted
in the child when we give him, and when he understands, the
definition. We need an account of the capacity that a person who
possesses the concept has, where an account of the capacity will
include a description of the background of human activity which
is the context for its acquisition and exercise.4 The definition of a
ruler is ‘measuring instrument’, so for a philosopher to explain the
meaning of ‘measuring’ in terms of the meaning of ‘ruler’ would

‘THE SUBJECT UNDER THREE HEADS’60

be ‘moving in a circle in our explanations’ (see contents §5, p. iii).
A philosophical account of the distinction, for Anscombe, should
start by describing the activities into which the use of these words
is interwoven; it should describe the background against which
measuring and estimating take place. The idea is that it is by
describing that background, and not by looking at what goes on
in particular cases of measuring and estimating, that we come to
understand the distinction that we draw when we say ‘That was an
estimate and not a measurement’.

The change in method can be summed up in terms of differ-
ing attitudes toward the following question: If I want to get clear
about what ‘intention’ means where should I look? For a philos-
opher engaged in answering Anscombe’s Question, the question
Anscombe is sometimes supposed to be addressing, the answer is:
we should look more closely at the phenomena to which ‘inten-
tion’ refers, the alleged states of mind and bodily movements. For
a philosopher like Wittgenstein – and Anscombe – we should look
not at the phenomena but at the linguistic practices and activities
in which it comes to be that ‘intention’ refers to those phenom-
ena. This is hard to do, because it seems to take our attention in
the wrong direction, but the thought is that once we understand
what it is we do when we use the word ‘intention’, we will come
to understand everything that is essential to the phenomena we
thereby describe.

For Anscombe, as for Wittgenstein, to describe a concept’s
character is to describe the look of a human life containing that
concept. For a concept as central to human life as intention, the
activities into which the concept is ‘interwoven’ will be manifold,
complex and multi-layered. The existence of the concept will
imply all sorts of facts about human physiology, environment and
behaviour, and about ‘the natural history of man’ including his
institutions, culture and society. The application and mastery of
the concept will require and imply all sorts of capacities, linguistic
and otherwise, and all sorts of experience and knowledge on the
part of the speakers. An account of the concept would show us
what our lives would look like if they did not include the concept
of intention, or if they included a concept that was akin to the
concept of intention but, for example, in which we did not express

‘THE SUBJECT UNDER THREE HEADS’ 61

intentions for the future. It would show us how our concept of the
involuntary, the voluntary, the unintentional, depend on, or relate
to, our concept of intention. It would describe how our practices
of giving orders, making predictions, explaining, blaming, hold-
ing responsible, are connected with our use of that concept. It
would say which of our concepts and practices depend for their
meaning or existence on the concept of intention. It would enable
us to see which primitive – non-linguistic, animal – capacities and
behaviours are connected with our concept of intention. I suggest
that Intention does all of this. It would, however, be the work of a
lifetime to lay all this bare; this Guidebook is just an outline.

Wittgenstein said: ‘We are not at all prepared for the task of
describing the use of e.g. the word “to think”’ (Wittgenstein,
1967, §111). So too, we are not at all prepared for the task of
describing the use of the word ‘to intend’. Part of what makes
Anscombe’s writings so exhausting is that she works constantly to
draw our attention away from the phenomena – away from how
it might feel or seem or look when we or others act intention-
ally or form an intention to act – and toward the ‘“enormously
complex tacit conventions” that accompany our understanding of
ordinary language’ (§43, p. 80) and against which it makes sense
to describe what is happening as intentional, or an utterance as
an expression of intention.5 Drawing this background to our
attention is unsettling, because as it comes into view a distinction
we found ‘intuitive’ starts to look like something ‘enormously
complicated’. ‘[A] child can give such a report!’ and yet once we
start to say ‘what is really involved in it’ this seems incredible
(§43, p. 80). Learning a concept like intention is ‘enormously
complicated’, not because a child must learn a set of rules, which
she then, mysteriously, forgets, but because her behaviour and
interaction with the world and with other humans must reach
a level of complexity and sophistication before the concept can
get traction. Think, for example, of everything a child must
understand and be able to do before the possibility that he owes
the grocer for potatoes can make sense (cf Wittgenstein, 1958,
§120). A human being handing potatoes to another human
being, no matter what is going on in their minds as they do so,

‘THE SUBJECT UNDER THREE HEADS’62

is only delivering potatoes that were ordered when each has acquired
certain capacities and understanding (e.g. to recognise money,
to understand the nature of exchange) against the background
of an enormously complex set of social and historical conven-
tions. Similarly, to ask of a baby’s movements whether they are
intentional or voluntary is at best poetic.

Part of what it is to have the concept is for one not to find its
application complicated. As Anscombe says in §4:

I am sitting in a chair writing, and anyone grown to the age of reason
in the same world would know this as soon as he saw me … if this
were something he arrived at with difficulty … then communication
between us would be rather severely impaired.

(§4, p. 8)

To make explicit the ‘enormously complex tacit conventions’
that are the background to his saying, straight off, ‘She is sitting
in a chair writing’, is to make what is simple complex. This is
one explanation of why one so often feels at sea when reading
Anscombe’s work.

It is for all these reasons that having Anscombe’s method in
view is crucial as one begins. The difficulty of her book is that
the description Anscombe is offering is one that seeks to make
explicit what we are already doing, happily, straight off, without
difficulty, whenever we use the concept of intention. But, we
want to object, if we are already making these distinctions, why
bother making explicit the rules we are already following? Surely
the important work is to look at the states of mind and the bodily
movements that are picked out, and to investigate their nature.
If one approaches Intention in this way, then not only will the
book be frustrating – much of what Anscombe says will seem
banal, irrelevant, or preparatory – but one will also miss the very
description she is offering. This will happen because if the work of
Intention is taken for non-essential ground clearing, a ‘first draft’
as that early reviewer put it, one will skim over it in the hope
of discovering the real thesis, hidden by all the digressions and
roundabout discussion. Such a search will be in vain.

‘THE SUBJECT UNDER THREE HEADS’ 63

WHAT IS THIS DESCRIPTION FOR?

Above I quoted the beginning of a remark from Zettel; here is the
rest of it:

We are not at all prepared for the task of describing the use of e.g. the
word ‘to think’. (And why should we be? What is such a description
useful for?)

(Wittgenstein, 1967, §111)

This brings out two dimensions of difficulty for a reader of
Intention, both of which I have already alluded to. We are not pre-
pared for the task Anscombe undertakes; and it is hard to see what
such a description is for. I want to say something about the second
of these now, before we prepare ourselves for the task Anscombe
undertakes.

If you ask me ‘What is a measurement?’ and you were hoping
for an answer by a few examples, then a lengthy and complex
description of the look of a life containing the concept of ‘meas-
urement’, one which attempted to draw from scratch a boundary
around all those bits of activity that count as measuring, even
though we can draw that boundary ourselves ‘straight off ’, will
surely test your patience. What could such a description be for?
Here are two answers.

When we think about ‘intention’ we are ‘tempted to speak of
“different senses” of a word which is clearly not equivocal’ (§1, p. 1).
We succumb to that temptation when we take Anscombe’s
Question at face value. That we are so tempted is a sign that we
are making certain assumptions that are obscuring a clear view of
its character. If we can uncover those assumptions, and give them
up, then we will no longer find ourselves on the edge of a con-
tradiction: ‘intention’ cannot be univocal, yet it is univocal. It is
better for our philosophical health if we can get this kind of ‘peace’
and, as Austin puts it at the end of his paper on pretending: truth
is important, and the task of classifying and clarifying is one way
of getting to the truth (Austin, 1958, p. 278). It will also save us
a lot of work, because we will no longer try to answer a question
which, it turns out, cannot be answered.

‘THE SUBJECT UNDER THREE HEADS’64

To look ahead, as we go through Intention we will see Anscombe
uncovering a whole web of assumptions that she recognises are
‘deep rooted’ (§2, p. 4) and which stand in the way of our being
able properly to recognise the unity of the concept of intention.
By ‘deep rooted’ she means: they are not connected with ‘super-
ficial grammar’ (§2, p. 4). An error connected with superficial
grammar would be one that arose simply from taking sentential
structure as a guide to propositional form. As if, for example, I
was confused by the sentence ‘No-one is coming’ into thinking
that someone must be coming because the sentence’s grammatical
subject – ‘No-one’ – must correspond to some entity of which
‘is coming’ is predicated. The superficial grammar of ‘No-one is
coming’ could be a source of error, and it is certainly helpful to
have these potential traps pointed out – not least because super-
ficial errors can lead to deeper ones – but they are not the ones
with which Anscombe is primarily concerned in Intention. Rather,
the mistakes Anscombe is concerned to uncover relate to what
Wittgenstein called depth grammar, which ‘includes activities other
than the production of language, into which a use of language
is interwoven’.6 These mistakes are connected with assumptions
we make about how our thought and our actions relate; the way
in which descriptions of events are answerable to the reality they
describe; and about what we can know and how. At the heart is
the following question: if an action is a physical happening, what
bearing can what the agent thinks have on the description of the
action? (Anscombe, 1969a, p. 10).

As Anscombe asks it there, we can see that there is a strong
temptation to answer: no bearing. If an action is a physical hap-
pening, then we ‘want to say that a physical happening is what
takes place whatever the agent thinks’ (Anscombe, 1969, p. 10).
Our objection to the idea that ‘what an agent thinks’ can have
a bearing on ‘the description of the action’ is clearly not one
that is rooted in superficial grammar, but is rather connected to
the way in which we think of ourselves, our judgements, our
actions, and the reality that surrounds us. To say, however, that
they are not rooted in superficial grammar is not to deny that
they are, ultimately, conceptual – or as Wittgenstein would say,
grammatical – problems.

‘THE SUBJECT UNDER THREE HEADS’ 65

We might ask the question again: why undertake the task of
recovering to view the character of the concept when we are per-
fectly at home using it. A non-philosopher might ask this, or at least
someone untroubled by the contradiction of seeing that the concept
is univocal but suspecting it cannot be. Intention is a lot of hard work
just for an abstract pursuit of Truth; counting the hairs on your head
by pulling them out one by one is another way of getting to the
truth, but not worth the pain or the effort. Of course, the truths that
Anscombe is offering are rather deeper, but one might remain unim-
pressed nevertheless. Here Anscombe has an answer which sets her
quite apart from Wittgenstein and recalls to mind the way in which
her philosophy is shaped by Aristotle and Aquinas, and the ‘righteous
fury’ (ERP, p. vii) of her political and religious writing. Anscombe
is not after ‘philosophical peace’. Rather, she thinks that it is only by
making it possible to see how and that ‘intention’ is univocal that it
will be possible to do the sort of moral philosophy which is capable
of ‘revolt[ing] against the conventional standards’ of middle class
thinking and ‘say[ing] something profound’ (1958b, p. 186). This
gives a different kind of importance, and a different kind of intelli-
gibility, to the project of grammatical enquiry to the one we find in
Wittgenstein’s work. For Wittgenstein, a good piece of philosophy
was one that allowed him to stop (Wittgenstein, 1958, §138); for
Anscombe it is one that allows her to show why vicious acts – bomb-
ing civilian populations, torturing a hostage, locking up the innocent
(as well as abortion, voluntary euthanasia, contraception, sex out-
side marriage) – cannot be made ‘morally good’ on grounds of
circumstance or effect. As I will present things, discharging the three
tasks of conceptual clarification required for moral philosophy to be
possible and describing the unity of the concept of intention are the
very same project.

‘A TOOL FOR THE PHILOSOPHY OF ACTION’

As we have already noted, one of the things that is disorienting
for a reader of Intention is Anscombe’s seemingly arbitrary atti-
tude when it comes to defending or arguing for the claims that
she makes. For example, the claim that a person knows what he
is doing without observation is introduced without argument or

‘THE SUBJECT UNDER THREE HEADS’66

explanation in §4 (p. 8) and then repeated up until §28, at which
point Anscombe throws everything into confusion by suggesting
that perhaps to say this is not ‘justifiable’ (§28, p. 49). Or, she says,
also in §4, that ‘if you want to say at least some true things about
a man’s intentions, you will have a strong chance of success if you
mention what he did or is doing’ (p. 8), but she doesn’t explain
how this can be so. On the other hand, when the suggestion is
made that intentional actions are ones that are not involuntary she
decries this as ‘question-begging’. The definition of an expression
of intention as ‘a description of something future in which the
speaker is some sort of agent, which description he justifies (if he
does justify it) by reasons why it would be useful or attractive if
the thing came true’ (§3, p. 6) is said to be ‘not false but rather
mystifying’ (§3, p. 7). Often what seems like the most contro-
versial of points is introduced as if it were a truism, and the least
contentious is rejected as ‘intuitive’ but false. If we are to attempt a
reading of Intention on which it is a sustained account of its topic,
rather than a loose collection of aphorisms, we must find a pattern
in this seemingly erratic behaviour.

A helpful place to start is with what Anscombe herself says about
one of those crucial statements which she nevertheless introduces
without defence of argument. It is that ‘an action may be inten-
tional under one description and not intentional under another’
(Anscombe, 1979b, p. 210; see §6, p. 11). If you ask me ‘Are you
doing that intentionally?’, I need to know what description ‘that’
is a placeholder for before I can answer. I may answer ‘yes’ if the
description is ‘signing a piece of paper’ but ‘no’ if it is ‘scratch-
ing the mahogany table-top’. So, my action is intentional under
the description ‘signing my name’ but not under the description
‘scratching the mahogany table-top’.

When Anscombe says that ‘an action may be intentional under
one description and not intentional under another’ (Anscombe,
1979b, p. 210) it can seem that she is putting forward a thesis
which might be debated or denied. It seems quite intelligible
(indeed, I’m sure it has happened) that someone may come along
and say, ‘You are quite wrong. Either an action is intentional
under every description or under none’. Or, ‘That is false, for the

‘THE SUBJECT UNDER THREE HEADS’ 67

description of an action is in no way relevant to the question of
whether it is intentional’. Read as a thesis that might be denied,
one expects to find in Intention an argument in its favour. But
while Intention contains no such argument, the whole book is
premised on its being true.

Writing in 1979 – with a view to rectifying ‘some of the
misunderstandings that [she has] noted’ when philosophers have
discussed her phrase ‘under a description’ (Anscombe, 1979b,
p. 208) – Anscombe says that some people have supposed that
when she says ‘one and the same action (or other event) may
have many descriptions’ that this is ‘said in the light of a theory
of event- identity’ (Anscombe, 1979b, p. 210). Such people have
thought that she holds that an action may be intentional under one
description and not intentional under another because she is com-
mitted to a prior thesis about the identity conditions for events. As
Anscombe no-where provides such a theory, it will be up to those
who wish to interpret or defend her account of intention to provide
one. Indeed, one of the reasons that Anscombe’s account of action
was subsumed by Davidson’s, is that Davidson does indeed provide
such a theory.

In ‘Under a Description’ Anscombe makes clear that she does
not mean §6 to be read in the manner I have just described.

When I introduced the phrase ‘under the description’ as a tool in the
philosophy of action [in Intention §6, p. 11], I thought it something that
couldn’t be called in question or misunderstood.

(Anscombe, 1979b, p. 208)

She meant, she says, ‘something that isn’t a philosophical the-
sis at all, and which no one denies’ (Anscombe, 1979b, p. 211).
Obviously, claims about action identity are philosophical theses
which people do deny. So what can Anscombe mean?

Anscombe compares the relationship between the statement:
‘one and the same action (or other event) may have many descrip-
tions’ and a theory of action- or event-identity, with that between
the statement: ‘one and the same man may satisfy many different
definite descriptions’ and a theory of human-identity (Anscombe,
1979b, p. 210). Different theories of human identity, she says, may

‘THE SUBJECT UNDER THREE HEADS’68

give different answers to the question ‘Is this the same human?’
in certain fictitious, unusual or borderline cases. For example, a
psychological theory may yield a different answer to a somatic
theory in cases involving brain transplants. Or, an account in
terms of memory might answer a question about the identity of a
person with Alzheimer’s differently to an account that grounded
personal identity in the identity of a human animal. However,
theories that differ on fictitious, unusual or borderline cases must
nevertheless agree on the vast majority of ordinary cases if they are
to count as theories about the same topic. A theory that denied
that ‘one and the same man may satisfy many different definite
descriptions’ would not, Anscombe says, even count as a theory
of human identity. Whatever such a theory was an account of it
would not be helpful in answering questions of human identity:

But what would we say of a theory [of human identity] which grants
that a certain man, Dickens, wrote David Copperfield and Bleak House
and that only this Dickens wrote David Copperfield, and only this
Dickens wrote Bleak House – but does not grant that ‘The author of
David Copperfield’ describes the same man as ‘The author of Bleak
House’ …? We’d say that it is a non-starter: any theory of human iden-
tity has got to fit in with the correctness of calling the author of David
Copperfield the same man as the author of Bleak House (subject to
astonishing literary discoveries, which are not our concern here). To
say that one must have a theory of human identity if one says that, or
that one needs it in order to justify saying that, seems to be absurd.
A theory, I suppose, will at least (a) determine answers in obscure
or borderline cases; (b) give an interpretation of known facts. It may
sometimes also correct generally received statements in the light of
new knowledge of a general scientific nature. The Dickens case is
affectable only under (b), i.e. not in such a way as to result in a change
in the truth-value we assign to the identity statement.

For this reason I always balked at the question ‘What is your theory of
event-identity?’ or ‘What theory of event-identity lies behind saying that
(in the imagined case) putting the book down on the table and putting
it down on an ink puddle were the same action?’ Any ‘theory of event-
identity’ had better yield this result: it itself is not a theory or part of one.

(Anscombe, 1979b, p. 210)

‘THE SUBJECT UNDER THREE HEADS’ 69

For Anscombe, then, any account of action which implied that a
single action does not have many different descriptions, or which
denied that a man may know he is doing a thing under one
description and not under another (just as a man may know him-
self under one description and not under another), or which did
not imply that an action may be intentional under the description
‘landing on a twig’ and not under the description ‘landing on
bird-lime’ (Anscombe, 1979b, p. 210), or which yielded the result
that ‘sawing Smith’s plank’ could be a description under which
Jones acted even though Jones didn’t know he was sawing Smith’s
plank, will simply not count as giving an account of action or
intentional action. Rather than proposing a thesis about inten-
tional action, §6 provides a constraint on an account of intention.

We see these statements explicitly playing this role in §20.
Anscombe rules out an account of intention as a ‘style characteris-
tic of observable human proceedings’ on the grounds that it does
not yield the result that an action may be intentional under one
description and not under another. ‘It is clear’, she says, ‘that a
concept for which this does not hold is not the concept of inten-
tion’; and, we can add, any account of a concept for which this
does not hold is not an account of intention (§20, p. 30).

If we are to follow the argumentative line of Intention, we need
to attune to the difference between a statement with which
Anscombe introduces a constraint on an account of action, and
one in which she develops an account. We also need to be aware
of a third kind of statement: one which articulates an intuitive or
plausible-sounding picture which Anscombe means to reject. We
can think of these kinds of statement as corresponding to three
voices, which I will call: descriptive, intuitive and philosophical.

In the descriptive voice, Anscombe makes statements which
are to act as constraints on an account of intention, in the way
just described. These are descriptions of our linguistic practice,
of non-obscure, non-borderline cases that are characteristic of
the concept of intention. Note that these are not necessarily
contexts in which we would naturally use the word ‘intention’.
Anscombe points out, for example, that crossing the road is a
clear case of an intentional action, though it would often be odd

‘THE SUBJECT UNDER THREE HEADS’70

to say ‘He crossed the road intentionally’ (see §19, p. 29). These
statements are neither defended nor justified, and the content is
uncontroversial – at least outside a philosophy classroom.

Another important example of this kind of statement is §4.
Anscombe refers here to ‘the sort of things you would say in a law
court if you were a witness and were asked what a man was doing
when you saw him’, and goes on:

[I]n a very large number of cases, your selection from the immense
variety of true statements about him which you might make would
coincide with what he would say he was doing, perhaps even without
reflection, certainly without adverting to observation.

(p. 8)

This fact is not there to be defended or challenged; it is not open
to say ‘No, that’s wrong. In a very large number of cases your true
statements would not so coincide’. Rather, we are reminded of
what an account of intention must accommodate. If it seems to us
that this cannot be true, the correct response, by Anscombe’s lights,
is not to defend the facts (what, after all, could this mean?), but to
look at why one feels inclined or compelled to deny them.

This brings us to the next voice, which I called intuitive.
These interjections are of two kinds. They either give voice to a
‘plausible’ or ‘natural’ explanation of the facts introduced in the
descriptive voice; or they are objections to the descriptions them-
selves. Anscombe introduces this perspective with phrases such
as: ‘Now it can seem that …’; ‘It is natural to suppose …’; ‘It is
often said that …’; ‘It is natural to object that …’; ‘One may feel an
objection …’; ‘Someone might say …’. Interjections which give an
intuitive explanation and those which object to unobjectionable
fact are related, in Anscombe’s view, in the following way: what is
leading the interlocutor to object that a description ‘which no one
can deny’ cannot be true (to deny the undeniable) is his adherence
to a ‘natural’ or ‘intuitive’ explanation of those facts. His intuitive
picture forces him to interpret the facts in a particular way and,
when he does so, contradictions arise.

Finally, there is Anscombe’s philosophical voice, introduced by:
‘We can clarify …’; ‘To understand …’; ‘What is required is …’;

‘THE SUBJECT UNDER THREE HEADS’ 71

‘We should look more closely at …’. Here Anscombe seeks to
undermine those intuitive pictures which lead the interlocutor
to deny what cannot be denied. This is done by describing the
character of the concept of intention.

In fact, in §1 we can see in miniature the way in which moving
between these three voices is to work in Intention. Anscombe
begins by describing what we say and do; for example, ‘Very often,
when a man says “I am going to do such-and-such”, we should
say that this was an expression of intention’. She points out that
if we set out to explain the phenomena in a way that is natural
or intuitive, but which is insufficiently attentive to the concept’s
overall use, we can end up saying something which is puzzling or
obscure, or which is in tension with the description with which
we began. Anscombe then proposes that we undertake a philo-
sophical enquiry which will replace what ‘we may be tempted to
think’ with a detailed description of the character of the concept
of intention.

Throughout this Guidebook I will draw attention to these voices
as they interact. It is often helpful, if one is stuck on a particularly
knotty paragraph in Intention, to ask: which voice is this?

Of course, there is question of priority here. Might the interlocutor
not protest as follows: surely a good, intuitive, natural explanation
which is inconsistent with one of Anscombe’s ‘descriptions’ can-
not be dismissed out of hand? On what ground does Anscombe
think her ‘descriptions’ can play the role of constraint in this way?

This is a difficult and important objection for any philosopher
who seeks to describe rather than theorise, but it is worth noting
that the way in which Anscombe appeals to facts which ‘no-one
can deny’ makes her resistant to two of the most common forms
of this objection. First, an ordinary language philosopher, like
Austin, prioritises what ordinary speakers say about and in lan-
guage. But one might well worry that the fact that we would find
it unnatural to use a particular word or expression in a situation
is not necessarily a good guide to the question of whether the
situation is properly characterised using a concept that the word
represents. Anscombe is not open to this objection. It ‘sounds
odd’ to say ‘He crossed the road intentionally’ but, Anscombe

‘THE SUBJECT UNDER THREE HEADS’72

points out, ‘it would be wrong to infer from this that we ought
not to give such an action as a typical example of intentional
action’ (§19, p. 29). For her, part of what it is to get clear about
the concept of intention is to understand why it sounds odd,
despite the fact, which ‘no-one can deny’, that crossing the
road is a very ordinary sort of intentional action. The facts that
Anscombe presents are not narrowly linguistic facts, though they
are facts about what we say and do. I do not offer that as a full
account of their status; this is something we will need to develop
as we work through Intention.

An objection of a different kind might be raised against a
philosopher like John McDowell. For McDowell, the aim of phi-
losophy is to provide a theoretical framework to shore up ‘what
comes naturally to pre-philosophically uncontaminated common
sense’ (McDowell, 1991, p. 298). There is certainly a question
of priority to be raised here: why should our pre- philosophical
intuitions be preferred to those shaped by philosophical under-
standing? Who says ‘uncontaminated common sense’ isn’t just
untutored prejudice? But Anscombe is attacking, rather than
defending, ‘what comes naturally’. She is not presenting views
that we find intuitively plausible but then give up in the face of
philosophical pressure. Rather, she is describing us doing things,
saying things, going about our lives, a description which we
give up in the face of pressure from ‘what comes naturally’. It
is harder to see how those descriptions might not be legitimately
taken as constraints.

THE STRUCTURE OF INTENTION

Now let me say something specific about the structure of Intention
and the way in which the descriptive, intuitive and philosophical
voices interact. This overview is necessarily dense and some of the
ideas it introduced will not become clear until much later; I rec-
ommend that a reader return to it when help seeing the wood for
the trees is needed.

In the opening paragraphs of Intention (§§2–4) statements are
mostly descriptive and intuitive and it is the tension between these
two perspectives which generates a series of difficulties and puzzles.

‘THE SUBJECT UNDER THREE HEADS’ 73

Anscombe begins with the fact (descriptive) that we sometimes call
‘I am going to do such-and-such’ an expression of intention and
sometimes a prediction, and that we call it the former when it is
justified by reasons why it would be good if the description came
true, and the latter if it is justified for reasons for believing it will
come true. Her interlocutor objects (intuitive) to calling an expres-
sion of intention a prediction, on the grounds that an expression
of intention for the future is not called ‘true’ or ‘false’ in the face of
what happens. The interlocutor also outlines a ‘natural’ picture on
which expressions of intention are descriptions or expression of ‘a
present state of mind’ (§2, p. 2). Anscombe finds that she cannot
get clarity about the cases she describes – that is, cannot develop
a philosophical account – so moves on to describe another set of
cases: those involving intentional actions. At the very end of the
book she returns to expressions of intention, in order to provide
the philosophical account which remained allusive in §2.

The discussion of intentional actions runs from §4 to §21 and
is largely a description of the facts that are to act on a constraint
on any philosophical account of intention. Here Anscombe seeks
to set out the distinctions that we make when we say that what
a person is doing either is or is not intentional. She does this by
describing when an answer to the question ‘Why?’ would assign
the description to a particular class. The interlocutor appears rarely
in these sections, and when he does it is usually to insist that a dis-
tinction that Anscombe is working away at is ‘intuitively obvious’
or ‘clear’; Anscombe is always insisting that it is not and often
claiming further that our taking it to be obvious or clear leads to
an interiorising of intention. It is in this context that Anscombe
rejects as ‘circular’ the idea that we can define intentional action
simply as not involuntary. What is difficult about these sections is
that Anscombe begins with a distinction which looks simple and
describes it in a way that makes it incredibly complicated.

§§22–44 come under Anscombe’s second head, intention with
which. These sections continue the work of §§4–21, and begin
to ‘reveal the order that is there in the chaos’ (§43, p. 80) that
those earlier paragraphs lay bare. The pattern that emerges in the
descriptions under which a man is doing what he is doing is, says
Anscombe a ‘calculative order’. This pattern reveals the unity of

‘THE SUBJECT UNDER THREE HEADS’74

the three ‘heads’ of intention. Here we begin to see the philosoph-
ical account that Anscombe is offering.

A major interjection is made in these passages by the interloc-
utor. Up until §28, a statement introduced in §4 the descriptive
voice – as a constraint on a philosophical account – has gone
unchallenged, but here the interlocutor insists, finally, that the
description cannot be true. This passage is one of the most well-
known in Intention:

‘Known without observation’ may very well be a justifiable formula for
knowledge of the position and movement of one’s limbs, but you have
spoken of all intentional actions as falling under this concept. Now
it may be e.g. that one paints a wall yellow, meaning to do so. But is
it reasonable to say that one ‘knows without observation’ that one
is painting a wall yellow? And similarly for all sorts of actions: any
actions that is, that are described under any aspect beyond that of
bodily movement.

(§28, p. 50)

The remainder of Intention (§§28–50) – excluding those final
two paragraphs on expressions of intention for the future – can
be viewed, in a large part, as in the philosophical voice: here
Anscombe attempts to diagnose and dislodge the knot of assump-
tions that lead to this interjection (and, as it turns out, also led
to the interlocutor’s resistance to calling expressions of intention
for the future a species of predictions). Her account of practical
reason provides most of the material to respond to this challenge
(§33, p. 57). The obstacle, Anscombe believes, comes in the form
of assumptions about the ways in which what a man knows or
thinks about what is happening can bear on what is happening –
assumptions which do not belong to superficial grammar. She
presents an account of practical reasoning which shows (a) that it
is essentially calculative, (b) that its conclusions are actions under
a description, and not propositions about what one ought to do,
(c) that it does not involve logical or psycho-causal compulsion.
She thinks that equipped with such an account there is no impedi-
ment to saying that when I know what I am doing it is not because
my description of what is happening matches reality, but because

‘THE SUBJECT UNDER THREE HEADS’ 75

reality matches my description of what I am doing. In this sense,
what I think can affect how what is happening can be described.

Once the impediment to accepting that has been removed, the
philosophical account of the concept offered between §4 and §28
can be reaffirmed and the implications drawn out.

A description of Intention’s overall shape, one which I will come
back to throughout this Guidebook, is in terms of a movement in
our understanding. At the start of Intention we are under the illu-
sion that character of the concept of intention is simple, and that
the difficult thing is to describe the property to which it refers.
Anscombe shows us that the concept is incredibly complicated by
outlining what seems like a chaotic ‘class of intentional actions’
(§§5–19). She then imposes order on this chaos via a description
of the formal character of practical reason (§§20–42). At the end
of the book (§§42–49) she explains why it is that, from the per-
spective of someone who answering the question ‘What is he / are
you doing?’ all that complexity and structure is quite hidden from
view. To map this onto the book’s chapters: Chapter 4 moves from
simplicity to chaos, Chapter 5 from chaos to order, Chapter 6
from order to simplicity.

NOTES

1 Throughout Intention, Anscombe uses ‘a man’ and the male pronoun in
her examples. Sometimes this makes the protagonist male – as when a
man is poisoning a well or going shopping – and sometimes it is equiva-
lent to the German word Mensch. Nowadays it would be more common, to
have female protagonists and to use a neutral form like ‘human’, ‘person’,
or ‘someone’ for Mensch. In this Guidebook I will follow Anscombe’s use,
and stick on the whole to ‘a man’ or ‘he’. This is partly because that sort
of revisionism feels rather odd, but mostly for stylistic reasons. It helps
with the integration of quotes, and also, as Anscombe is a woman, avoids
ambiguity in anaphoric reference.

2 This label is coined by Moran and Stone (2011).
3 It is worth noting the deep connections between Anscombe’s Question

and what has come to be known as ‘Wittgenstein’s Question’. Wittgen-
stein asks, in Philosophical Investigations §621: ‘What is left over if I

‘THE SUBJECT UNDER THREE HEADS’76

subtract the fact that my arm goes up from the fact that I raise my arm?’
Many philosophers of action have taken this question at face value,
and have sought to solve the equation: arm rising + ? = arm raising. It
is clear from the context, however, that Wittgenstein’s point in raising
the question is to reveal what is at fault in a particular way of think-
ing about action. For an illuminating discussion of the relation between
Wittgenstein’s Question and Anscombe’s Intention see Ford (2011). See
also Chisholm’s review of Intention for an illustration of what happens
when one tries to read Intention as an answer to this equation.

4 Compare Geach (1958, p. 12). This, of course, is the background against
which one description – holding this stick against this object – can be
‘brute’ relative to another – measuring the object’s length.

5 The quote is from Wittgenstein Tractatus (1922, §4.002).
6 See Gordon Baker (2001).

SUGGESTED READING

Anscombe, Elizabeth (1979b). ‘“Under a Description”’. Nous, 13. Reprinted in MPM,
pp. 208–219.

Austin, J. L. (1957). ‘A Plea for Excuses’, Proceedings of the Aristotelian Society, vol. 57.
Baker, Gordon (2001). ‘Wittgenstein on “Depth Grammar”’. Reprinted in Katherine

J. Morris (ed.), Wittgenstein’s Method: Neglected Aspects. Essays on Wittgenstein by
Gordon Baker (Oxford: Blackwell, 2004), pp. 73–91.

McGinn, Marie (1997). Routledge Philosophy Guidebook to Wittgenstein’s Philosophical
Investigations (London: Routledge), chapter 1.

Moran, Richard and Stone, Martin J. (2011). ‘Anscombe on Expressions of Intention:
An Exegesis’. In Anton Ford, Jennifer Hornsby and Frederick Stoutland (eds),
Essays on Anscombe’s Intention (Cambridge, Mass.: Harvard University Press),
pp. 33–75.

4
INTENTIONAL ACTION

(§§5–19)

Anscombe opens her investigation into intentional action with a
simple statement: ‘Intentional actions are ones to which a certain
sense of the question “why?” has application’. What follows is her
attempt to ‘clarify’ that account (§6, p. 11). It is crucial to remember
that Anscombe’s method of clarification is not at all one of simpli-
fication. Rather, she begins with what seems simple, and gradually
uncovers its complexity. Consider again the complexity involved in
describing the background conditions in which someone leaving
potatoes at a door would be supplying me with potatoes.

These paragraphs seem very untidy; a series – or maybe an
unordered list – of distinctions and classifications, rather haphaz-
ardly presented. Anscombe steadfastly refuses to leave a distinction
unmade, even where doing so would make things cleaner and
more straightforward for the reader. In these paragraphs there is
no grand resolution or major revelation. Rather, there is a growing
appreciation of the ‘“enormously complicated tacit conventions”
that accompany our understanding of ordinary language’ (§43,
p. 80), and a deepening realisation that the distinction between

INTENTIONAL ACTION78

actions which are intentional and those that are voluntary or
involuntary cannot be ‘appealed to as something intuitively clear’
(§2, p. 1). It is not until §22 – the starting point of our next
chapter – that we begin to find some relief, as Anscombe starts to
describe an ‘order’ or pattern in the ‘chaos’ she has created (§43,
p. 80). It is in these passages that Wittgenstein’s advice is most apt:
‘Take your time!’

THE QUESTION ‘WHY?’ (§§5–18)

Anscombe begins her enquiry into the concept of intention
under its second head, intentional action, with the question
‘What distinguishes actions which are intentional from those
which are not?’. She immediately offers a simple answer: ‘they
are the actions to which a certain sense of the question “Why?”
is given application; the sense is of course that in which the
answer, if positive, gives a reason for acting’ (§5, p. 5). This
answer is in what I called the descriptive voice; nothing that fol-
lows should be taken to undermine or challenge it. She then
proposes to ‘clarify’ this answer by explaining the sense of the
question ‘Why?’. The ‘clarification’ takes place in eleven sections
between §5 and §18.

As we know, for Anscombe, clarification does not mean
simplification: rather, to clarify is to reveal and describe the
extraordinarily complex linguistic practice – where linguistic
practice, recall, includes ‘activities other than the production of
language, into which a use of language is interwoven’ – within
which an action can be described using the concept of intention.
The crucial thing to avoid, when seeking clarity of this kind, is
the ‘danger of moving in a circle in our explanation’ (contents,
§5, p. iii). We already saw when we discussed the concept of
‘length’, that an explanation of length in terms of measuring
would presuppose the linguistic practice which was the back-
ground for descriptions using the concept of length. Hence, it
would not count as an explanation by Anscombe’s lights. A circu-
lar explanation occludes a layer of complexity – takes something
for granted that needs to be laid bare – and hence undermines the

INTENTIONAL ACTION 79

task of clarification. Returning to the concept of intention, the
simple answer given in §5 is clearly circular because ‘the question
“What is the relevant sense of the question ‘Why?’” and “What
is meant by ‘reason for acting’?” are one and the same’ (§5, p. 9).
But in these passages a number of other, less obviously circular,
ways of explaining the sense are rejected on the same ground: for
example, an appeal to an intuitive understanding of the distinc-
tion between reason and cause (§5, p. 9–10), or to the concept of
the involuntary (§7, p. 12).

In order to avoid circularity, Anscombe proceeds by ‘roughly
outlin[ing] the area of intentional actions’ (§18, p. 28), aiming
to do so in a way that does not presuppose that the sense of the
question ‘Why?’ is already understood. If she succeeds, and if this
class is just the class to which the question ‘Why?’ applies, then in
outlining the class she will have explained the question’s sense by
describing its application.

She takes as her starting point the class of descriptions of what
a man is doing under which he knows he is doing it. This restric-
tion will be important later: for now we should just note it, and
note that no explanation is offered of why it is that the descriptions
we are interested in – descriptions of intentional actions – will be
a sub-class of it. Anscombe then methodically, with enormous
discipline, divides that class into (1) descriptions which are jus-
tified by observation (§8); (2) descriptions which are justified
by mental causes (§§9–11); (3) descriptions which are justi-
fied neither by mental cause nor by observation (§§12–15);
(4) descriptions which are of the sort to fall under (3) but which
are not justified at all (§§17–19). These classes correspond,
respectively to involuntary movements, involuntary actions, inten-
tional actions, voluntary actions.

To help guide the reader let me make some further brief comments
about the structure of these passages. The structure, it must be
admitted, is painfully difficult to discern.

After saying that intentional actions ‘are the actions to which a
certain sense of the question “Why?” is given application’ (§5, p. 5),
Anscombe states that she will:

INTENTIONAL ACTION80

both explain this sense and describe cases shewing [sic] the question
not to have application. I will do the second job in two stages because
what I say in the first stage of it will be of use in helping to explain the
relevant sense of the question ‘Why?’.

(§6, p. 11)

The two jobs yields the following answers, summarised by Ans combe
in §16:

A: When does the question ‘Why?’ lack application? (§§6–9
and §§17–18)

Answer: (first stage) ‘the question has not that sense if the
answer is evidence or states a cause, including a mental
cause’ (§16, p. 24) or, (second stage) sometimes, if the an-
swer is ‘I don’t know why I did it’ or ‘No reason’ (§§17–18).

B: What is the sense of ‘Why?’? (§§9–15)
The answer may (a) simply mention past history, (b) give
an interpretation of the action, (c) mention something
future. In cases (b) and (c) the answer is already chara-
cterised as a reason for acting, i.e. as an answer to the
question ‘Why?’ in the requisite sense; and in case (a) it is
an answer to that question if the ideas of good or harm are
involved in its meaning as an answer; or again if further
enquiry elicits that it is connected with ‘interpretative’
motive, or intention with which (§16, pp. 24–25).

When Anscombe remarks that what she says in the first stage of
her enquiry into when the question doesn’t apply (job A) will help
her to explain when the question does apply (job B) she means that
what she says in defining the notion of ‘cause’ which is opposed
to ‘reason’ (first stage of job A) will help her to define ‘reason for
acting’ (job B). This help takes two forms.

First, defining ‘cause’ in this context helps her to locate the
difference between answers under B(a) which give a reason for
acting, and those that give a mental cause. She says that an answer
which ‘simply mention[s] past history’ gives a reason for acting
(rather than a cause) if ‘the ideas of good or harm are involved

INTENTIONAL ACTION 81

in its meaning as an answer’ (§13, p. 18). Second, it enables her
to block a ‘very natural conception of “motive”’ as ‘what moves …
glossed as “what causes” a man’s actions etc.’ (§11, p. 18). This is
important because very often, to give a motive is to give the inten-
tion with which one acts. Here Anscombe’s real opposition to the
so-called causal account of action emerges.1

‘IT WAS INVOLUNTARY’ AND REASON
VS CAUSE (§§5–11)

Anscombe’s first job is to describe when the question ‘Why are
you X-ing?’ or ‘Why did you X?’, where ‘X’ is a description within
the restricted range – i.e. is a description under which he knows
what he is doing, and which he knows without observation – lacks
application. The central concern in stage one of this job – under
taken between §5 and §11 – is to find a way through a difficulty
that she introduces in §5:

[C]onsider the question ‘Why did you knock the cup off the table?’
answered by ‘I thought I saw a face at the window and it made me
jump’. Now, so far [in the discussion of expressions of intention] I
have only characterised reason for acting by opposing it to evidence
for supposing the thing will take place – but the ‘reason’ here was not
evidence that I was going to knock the cup off the table. Nor can we
say that since it mentions something previous to the action, this will
be a cause rather than a reason; for if you ask ‘Why did you kill him?’
the answer ‘He killed my father’ is surely a reason rather than a cause,
but what it mentions is previous to the action.

(§5, p. 9)

This difficulty relates, as I have explained, to B(a): when is an
answer which ‘simply mention[s] past history’ an answer to the
question ‘Why?’ and when is it an answer which gives a cause?
Again, what looked like a simple distinction is about to be made
incredibly complex.

As the quoted passage makes clear, Anscombe rejects an answer
in terms of an intuitive distinction between reason and cause; she
says that ‘the topic of causality is in too great confusion’ for such an

INTENTIONAL ACTION82

answer to be ‘enlightening’ (§5, p. 10). This may come as a surprise
to anyone who thinks that ‘Hume pretty well dealt with the topic of
causality’ (§10, p. 16) or who is persuaded by Davidson’s claim that
we can look to science to tell us what counts as causation (Davidson,
1978, p. 83). However, whatever the state of philosophical debate
about causation, it does seem clear that our intuitive understand-
ing of the distinction is not sufficient to anchor a philosophical
understanding of the distinction between reason and cause. As we
will see, Anscombe in fact thinks that ‘I thought I saw a face at the
window and it made me jump’ is a causal statement of a kind that
is ‘so far from accommodating itself to Hume’s explanations that
[Humean’s] would leave it out of their calculations’ (§10, p. 16).

The answer ‘I thought I saw a face at the window and it made
me jump’ characterises the action ‘Knocking the cup off the table’
as involuntary; so a second suggestion might be that actions which
are caused are involuntary while actions done for reasons are inten-
tional. However, Anscombe calls this ‘question-begging’ (contents,
§8, p. iv); the notion of the involuntary, she says, ‘cannot be intro-
duced without treating as solved the very kind of problem we are
discussing’ (contents, §7, p. iv).

Later, we will understand fully why Anscombe insists on this –
insists, that is, that the notion of the involuntary cannot yet be
introduced. However, we can get a rough idea of her reasons
by briefly looking back to the ethical project we described in
Chapter 2. This will be a way to bring out some of the complex-
ities concerning the concept ‘involuntary’.

The class of voluntary actions is extremely important to Catholic
moral theology and, more generally, to the anti-consequentialist
ethical project Anscombe describes in ‘Modern Moral Philosophy’.
Recall that we identified three tasks of conceptual clarification
Anscombe thought to be necessary before it would be ‘profitable …
to do moral philosophy’ (Anscombe, 1958b, p. 169). The first and
third tasks were, respectively: to describe the distinction between
an action’s intended and its merely foreseen consequences; and to
give an account of human action which could support the idea
that an action’s moral character (as virtuous or vicious) could be
related to the kind of action that it was. Both these tasks can each
be reconceived as questions about voluntary action.

INTENTIONAL ACTION 83

Relating to the first task, Anscombe says – toward the end of
Intention – that the ‘antecedently known concomitant result of one’s
intentional action’ can be called either ‘“voluntary” or “involuntary”
in contrast with “intentional”’. The fact ‘one could have prevented it
if one would have given up the action’ makes such a result ‘ voluntary’;
however, if ‘one regrets them very much’ this may lead to the result
being called ‘involuntary’ (§49, p. 89). (This is another point where
an ‘intuitively clear’ distinction turns out to have a character quite
different from what we might have expected.) For Anscombe, then,
the first task can be recast as one about when a result of one’s action
is voluntary or involuntary, and when intentional.

Relating to the third task, and using a different sense of ‘vol-
untary’, one which contrasts with what I will here call involuntary
movements, Anscombe says ‘human action = voluntary action’.2 She
describes ‘digesting your food and breathing and sweating’ as well
as ‘idly stroking your beard’ as ‘the acts of a human agent but not
what I call human actions’ (Anscombe, undated a, p. 203; 1982b,
pp. 208–209).3 She proposes the thesis: ‘all human action is moral
action. It is all either good or bad. It may be both’ (Anscombe,
1982b, pp. 208–209).

As this brief digression makes clear, the concept of the involun-
tary, and the contrast that one seeks to make by calling what one
did ‘involuntary’, is both central to Anscombe’s ethical concerns,
and extraordinarily complex. The sense in which stroking my
beard is ‘involuntary’, and the sense in which doing you a favour
when I meant to harm you is ‘involuntary’, are clearly quite differ-
ent. As Anscombe puts it in §7, ‘[w]e can easily get confused by the
fact that “involuntary” neither means simply non-voluntary, nor
has an unproblematic sense of its own. In fact this pair of concepts
is altogether very confusing’ (§7, pp. 12–13). What comes out of
§§5–18 is ‘clarification’ – though by no means a simplification – of
these concepts.

KNOWN WITHOUT OBSERVATION (§8)

Anscombe identifies two kinds of answer to the question ‘Why
are you X-ing?’ or ‘Why did you X?’ which mention past history
and which give a cause rather than a reason. Both, she says, can be

INTENTIONAL ACTION84

described without ‘using any notions like “intended” or “willed”
or “voluntary” and “involuntary”’ (§8, p. 13) by ‘point[ing] out
a particular class of things which are true of a man: namely the
class of things which he knows without observation’ (§8, p. 13).
The concept of observational knowledge is one that can be defined
independently of this set of concepts, and so enables Anscombe
to describe two classes of involuntary action ‘without begging
any questions’ (contents, §8, p. iv). It is the second of those two
classes, illustrated with the example ‘He withdrew his hand in
a moment of involuntary recoil’ (§7, p. 13), that is the one for
which ‘we have difficulty describing the distinction between a
cause and a reason’ (§9, p. 15) – that is, the case that features in
the puzzle of §5.

Anscombe illustrates the class that she wants to introduce –
‘the class of things which [a man] knows without observation’ (§8,
p. 13) – with the example of the knowledge that a man has of the
position of his limbs. Such knowledge, she says, ‘is without obser-
vation, because nothing shews him the position of his limbs; it is
not as if he were going by a tingle in his knee which is the sign that
it is bent and not straight’ (§8, p. 13). Continuing her negative
definition, she says that:

Where we can speak of separately describable sensations, having
which is in some sense our criterion for saying something, then we
can speak of observing that thing; but that is not generally so when we
know the position of our limbs.

(§8, p. 13)

Anscombe does not dwell on the question of how one is able to
give the position of one’s limbs without relying on observation;
that capacity she says, is not her topic in Intention. By way of
further explanation, she remarks only that ‘the topic [of knowl-
edge of limb position] is certainly a difficult one, deserving of
fuller discussion; here, however, such a discussion would be out
of place’ (§28, p. 50).4 I will come back to why it is ‘out of place’
in a moment.

Though the question of how one knows the position of one’s
limbs is, Anscombe says, ‘out of place’ in Intention, the class of

INTENTIONAL ACTION 85

things known without observation is of central importance. This
is not just because it is part of Anscombe’s non-circular descrip-
tion of a pair of classes of involuntary actions. Anscombe says that,

the class of things known without observation is of general interest to
our enquiry because the class of intentional actions is a sub-class of it.
I have already said that ‘I was not aware I was doing that’ is a rejection
of the question ‘Why?’ whose sense we are trying to get at; here I can
further say ‘I knew I was doing that only because I observed it’ would
also be a rejection of it.

(§8, p. 14)

Before we look more carefully at what it might mean to say that
something is or is not ‘known by observation’, we need to consider
briefly the way in which this idea from Intention has taken on a
life of its own in some branches of anti-Cartesian philosophy of
mind. The idea that we know the position of our limbs and our
intentional actions ‘without observation’ has become a locus for
interpretative skirmish and philosophical debate.

Many have taken Anscombe to be putting forward a simple but
exciting thesis: the epistemological capacity that gives us knowl-
edge of the position of our limbs and our intentional actions
is of a kind with that which gives us knowledge of our mental
states. Epistemology often makes a distinction between self- and
other-knowledge; between the knowledge that I have of my own
states and dispositions, and the knowledge that I have of everything
else. The observational/non-observational dichotomy appears to
map onto this: my knowledge of myself is non-observational and
my knowledge of others is observational. This picture seems to
work well for the kinds of self-knowledge that are traditionally the
concern of philosophy of mind; we say: it is not by observation
that I know I am in pain, or that I know my beliefs, desires and
intentions. But the idea that this special kind of knowledge might
extend beyond the narrowly psychical – what Anscombe seems
to be suggesting when she gives limb position as an example of
knowledge without observation – seems to have profound impli-
cations for our understanding of the way in which the mind is
embodied. The possibility that this claim seems to be hinting at,

INTENTIONAL ACTION86

is that the self – the object of self-knowledge – can be identified with
‘human animal’ whose physical states are known in this privileged
manner, rather than with some part of that animal, the mind,
brain, or Ego. This, of course, is the prize at which anti-Cartesian
philosophy of mind aims.

Anscombe’s remarks about the nature of this special ‘non-
observational knowledge’ can amount to little more than a hint,
or the hint of a hint, of what this species of knowledge might
look like. Her characterisation, at least so far as bodily posture
goes, is wholly negative. But the prize is so great that philoso-
phers have been keen to pursue that hint as far as possible. Like
Fermat’s marginalia, the suggestion that someone had ‘discovered
a truly remarkable proof ’, a solution to the mind–body problem,
which the margin (in this case, the book) was too small to con-
tain, leads others to try reconstruction. But, as Godfrey Vesey
puts it, the ‘attempt to understand what Miss Anscombe means
by “knowledge without observation” is like a treasure hunt’ which,
unfortunately ‘seem[s] to lead, not to philosophical treasure at
all, but to … absurdity’ (Vesey, 1963, p. 201). The immediate
‘treasure’ is ‘an insight into intentional action’ (p. 209); the real
prize an insight into ‘the way in which the mind is embodied’
(p. 210). Philosophers on this treasure hunt are compelled to treat
Anscombe’s talk of ‘separately describable sensations’ as somehow
unlocking the notion of ‘non-observational knowledge’. This has
led to all sorts of odd conjecture, which I will briefly outline.

It has proved extremely difficult to produce a gloss on ‘separately
describable sensation’ which, when used as a criterion for observa-
tional knowledge, does not leave that account immediately open
to counter-example (see Martin, 1971, pp. 15–24). For example,
suppose we take Anscombe’s criterion of ‘separately describable
sensations’ to imply that with observational knowledge my descrip-
tion of how things appear mustn’t be a simple repetition of the
description that is used in the first person statement. This criterion
in terms of language runs aground on interpretative and substan-
tive issues. First, Anscombe’s own examples contradict this reading:
‘It felt as if there was fur’ is a repetition of the description ‘There is
fur’ (Anscombe 1964, pp. 39–40). Second, sometimes we have no
language for describing how things appear (other than by using a

INTENTIONAL ACTION 87

description of how things would be were they how they appeared)
due to paucity in our sensation language, especially for sense
modalities other than sight. Sometimes descriptions of how things
are, become part of our descriptive repertoire of appearances. This
is seen most clearly in a phrase like ‘the sensation of flying’, which
is a description of an appearance which makes use of a non-sensory
concept, ‘f lying’.5 For some descriptions, ‘of onions’ is an example
Anscombe gives, the description hovers between the two:

in ‘the smell of onions’ for example, ‘of onions’ is an external descrip-
tion, but English contains no word for the content. If onions ceased
to have that small, but it were still to be smelt elsewhere, ‘of onions’
might become an internal description, like ‘bitter’.

(Anscombe, 1962b, p. 72)

How could these truly contingent features of our language be the
Archimedean point on which to build a solution to the mind–
body problem?

Another suggestion sometimes made is that what Anscombe
means when she says that knowledge of the position of our
limbs is not by observation is that the knowledge is not inferred
from sensations – thus, talk of ‘separately describable sensations’
becomes talk of ‘sensations’ and ‘some sense criteria’ becomes
‘inferential ground’ and the part that we had such difficulty
unpacking is dropped. But then Anscombe’s claim is quite
consistent with the idea that the knowledge is perceptual –
perceptual knowledge is not inferential – in which case the label
‘non-observational’ is misleading at best (Pickard, 2004) and the
treasure vanishes. This would be disastrous so far as a reading of
Intention goes because, as we will see, it is central to Anscombe’s
view of the knowledge we have of our intentional actions that it
is not perceptual knowledge.

These difficulties have encouraged many to try to make sense
of non-observational knowledge without relying too much on the
idiosyncratic idea of separately describable sensations.

There is something simplistic about the way in which the
debates I have just outlined treat the contrasts between
observational and non-observational knowledge and between

INTENTIONAL ACTION88

self- and other-knowledge. The idea that these contrasts mark
neat dichotomies, dichotomies which also happen to map sim-
ply onto each other, fits uneasily with Anscombe’s philosophical
approach. Her concern is always to undermine and complicate
simple dichotomies, often by revealing that what looks like
straightforward disjunction is nothing of the kind. The idea that
an intuitive distinction between what we know with and with-
out observation would map, neatly, onto one between what we
know ‘as self ’ and ‘as other’ is alien to Anscombe’s way of think-
ing. If the reminder that we know some things not on the basis
of observation is to play a role in overcoming the myth of the
Cartesian subject – and Chapter 6 will suggest that it does – we
should not expect the route to be so direct. Anscombe offers
something much more complex, in which the remark about
‘separately describable sensations’ is one tiny piece.

Let us, for now, put that larger anti-Cartesian project from our
minds, and try to understand Anscombe’s remark that ‘where we
can speak of separately describable sensations, having which is in
some sense our criterion for saying something, then we can speak
of observing that thing’ (§8, p. 13). This remark, note, gives a
criterion for observation not for non-observation. Importantly,
it does not imply that knowledge of posture and knowledge of
intentional action are of the same kind, or that the capacity to
say how my limbs are arranged and the capacity to say what I
am doing are equivalent.6 Indeed, it will be crucial to recognise
that they are not. Our discussion here, then, is wholly negative –
we are aiming to describe only what Anscombe thinks is not the
case when we can say how our limbs are arranged or what we are
doing. However, it is nevertheless important, as we will need to
be clear about what is so excluded later, when we come to address
the topic of practical knowledge.

Anscombe wrote several papers on the concepts of sensation
and observation: ‘On Sensations of Position’ (1962b); ‘Substance’
(1964); ‘The Intentionality of Sensation: A Grammatical Feature’
(1965b); ‘Comments on Professor R. L. Gregory’s Paper on
Perception’ (1974e); ‘The Subjectivity of Sensation’ (1976b). As
with her investigation into intention, these discussions seek to

INTENTIONAL ACTION 89

describe, in detail, the linguistic practice associated with sensa-
tion concepts and the idea of observation or perception. As with
Intention, these discussions aim to make explicit the complexity in
what seems simple, and so resist easy summary.

What I want to point to is a discussion from ‘The Intentionality
of Sensation: A Grammatical Feature’ (1965b). In that paper
Anscombe seeks to arbitrate between an ordinary language
philosopher and a Berkleyan or Russellian ‘sense-data’ theorist
(1965b, p. 11). The dispute is about the immediate objects of
sight: are they ‘ordinary’ objects, like tables and chairs, or are they
‘peculiar’ objects like sense-data?

As the subtitle of her paper implies, Anscombe proposes a
grammatical solution to the dispute, one which she says shows
that both views are mistaken (1965b, p. 11). Both are mistaken
because both take the word ‘object’ in ‘object of sight’ to mean
‘thing’ or ‘entity’; what they disagree about is the sort of entity:
sense data or material. Anscombe argues that ‘object’ in ‘object of
sight’ does not mean entity because the category ‘object of sight’
is a grammatical rather than an ontological one. The question
‘What is the object of sight?’ is answered by giving the phrase
that follows the verb ‘to see’ in the sentence ‘I see such-and-such’.
Grammarians call this phrase the direct-object phrase. Anscombe
says that just as a debate about what kind of entity a direct object
is would be nonsense, so too is a debate about the ontological
status of objects of sight.

Anscombe says that taking the word ‘object’ in ‘object of
sight’ to be introducing a thing – albeit a non-material thing for
the sense-data theories – leads both the sense-data and ordinary
language philosopher to miss the fact that ‘verbs of sense-
perception … essentially have an intentional aspect’ (1965b,
p. 11). If some thing is seen then there is a description under
which it is seen, and that description is the direct-object phrase
in ‘I see such-and-such’. Not any description which is in fact
true of the thing can be the direct-object phrase, and so not any
description will give the object of sight. This goes for all inten-
tional verbs: the objects of knowledge, desire, belief, thought,
worship, etc., are not things. As Anscombe puts it, somewhat
gnomically, in Intention: ‘we must always remember that an

INTENTIONAL ACTION90

object is not what is aimed at [thought about, seen, wanted, etc.]
is; the description under which it is aimed at is that under which
it is called the object’ (§35, p. 66). An intentional object is thus
essentially related to a subject who is doing the seeing, knowing,
desiring and so forth.

The topic of intentional objects is a deep one that weaves its
way through Intention. Our focus here, however, is restricted to
the topic of objects of sensation. We wanted to explain Anscombe’s
remarks about the role of ‘separately describable sensations’ in
observational knowledge. I think her point is this simple one.
Sensation verbs take intentional objects, and as such if we speak of
observing something we imply the existence of a description under
which what is seen (heard, felt, etc.) is seen, a description which
would be the direct-object phrase in ‘I see/hear/feel such-and-
such’. For sensory verbs these descriptions give the look (sound,
feeling, etc.) of things.

To say that these sensations are ‘separately describable’ is to
recall the role that these descriptions play in cases of error. When
a person says, e.g. ‘There’s my father’, and his father is nowhere
to be seen, we can ask him to describe how things look, in that he
sees his father:

[I]f a man says he saw a man standing in a certain place, or heard
someone moving about, or felt an insect crawling over him, it is possi-
ble at least to ask whether he misjudged an appearance, a sound or a
feeling; that is, we can say: Look, isn’t this perhaps what you saw? and
reproduce visual effect of which he may be able to say ‘Yes, that is, or
could be, what I saw, and I admit I can’t be sure of more than that’;
and the same with the sound or the feeling.

(§28, p. 49)7

This is why, when we can ‘speak of observing’ we can also speak of
‘separately describable sensations’.

What then of knowledge of limb position? Anscombe says
that when someone erroneously gives the position of his limbs
‘it would be incorrect to say that he had misjudged an inner
kinaesthetic appearance as an appearance of his leg bent’ (§28,
p. 50). It is not that she thinks there is no such thing as making

INTENTIONAL ACTION 91

a mistake here – ‘if someone says that his leg is bent when it is
straight, this may be surprising but it is not particularly obscure’
(§8, p. 14) – but only that such an error does not imply that he
‘misjudged an appearance’. Rather, he thought his leg was bent
and it was not.

The argument about limb position will not be settled here. If we
grant Anscombe’s claim that such knowledge is non- observational,
in the way just described, then the question immediately arises:
well, how do I know, then? Anscombe does not answer this ques-
tion; that capacity, she says, is not her topic in Intention. However,
she does give us a clue as to what a capacity for knowledge of
observable phenomena – e.g. the states and motions of a par-
ticular human being – might look like if it is not grounded in
perception. This clue is buried in a rather curious comment about
the relationship between error and knowledge:

Where we can speak of separately describable sensations, having
which is in some sense our criterion for saying something, then we
can speak of observing that thing; but that is not generally so when we
know the position of our limbs. Yet, without prompting, we can say it.
I say however that we know it and not merely can say it, because there
is a possibility of being right or wrong: there is a point in speaking of
knowledge only where a contrast exists between ‘he knows’ and ‘he
(merely) thinks he knows’.

(§8, p. 14)

Michael Thompson dismisses this remark as a ‘peculiar obses-
sive theoretic tic … characteristic of Anscombe’s teacher Ludwig
Wittgenstein’. He finds no reason to hold to a ‘dogma about
the relation of knowledge to narrowly psychical phenom-
ena like pain or belief and intention itself ’ (Thompson, 2011,
p. 198). He then translates the ‘dogma’ into a gesture toward a ‘sui
generis’ species of knowledge (Thompson, 2011, p. 198–199).
But Anscombe is not being dogmatic here. That is to mistake a
description for a theory. What she says is that there is ‘a point in
speaking of knowledge’ (emphasis added) only where there is a
contrast between thinks and knows. Of course, one may speak in

INTENTIONAL ACTION92

a way that has no point if one wishes; thus one can certainly speak
of knowing where one’s pain is. But, if knowing amounts to no
more than being able to say, then it is better not to because speak-
ing of knowledge in such a context might lead one to think one
was making a distinction that did not exist, e.g. might lead one to
think that knowing where it hurt needed to be differentiated from
being able to say where it hurt. This then might lead one to specify
a ‘sui generis’ species of knowledge.

When I said that this gives us a clue about what a capacity
for knowledge without observation might look like, the feature
I wanted to draw attention to was the following: Anscombe says
that we have a ‘capacity to say’ how our limbs are arranged, and
that we can speak of knowing because there is a possibility of error.
It is this structure – capacity to say plus possibility of error – that
appears again when it comes to knowledge of what I am doing.
But though the structure is the same, the content of the capacity
to say is quite different.

What Anscombe doesn’t develop in Intention is any account
of what that capacity to say in relation to limb position amounts
to; I would suggest that it is a linguistic capacity intimately con-
nected with the capacity for voluntary movement. Later, we will
see that Anscombe’s account of practical reasoning can be thought
of as an account of a capacity to say which is intimately connected
with the capacity to do. Again, we will find ‘a point of speaking of
knowledge’ because there is a possibility of error.

INVOLUNTARY MOVEMENTS (§§7–11)

Let us now return to our main thread. Job A, recall, was to describe
when the question ‘Why?’ lacks application. The first answer was:
it lacks application when it gives a cause rather than a reason. But
Anscombe thought this answer was inadequate because the notion
of ‘cause’ is here unclear. The case illustrated in §5 was meant to
show this. The second answer was: it lacks application when it
is ‘It was involuntary’. But Anscombe thought this answer was
inadequate because the notion of ‘involuntary’ is both unclear and
‘cannot be introduced without treating as solved the very problem
we are discussing’ (contents, §7, p. iv).

INTENTIONAL ACTION 93

The class of things a man knows without observation is intro-
duced as a way out of this difficulty. With it to hand, Anscombe
can describe without circularity two kinds of answer to the ques-
tion ‘Why?’ which mention past history, but which give a cause
rather than a reason. The answers correspond to two classes of
the involuntary.

To the first class belong what we above identified as a class of
‘acts of a human agent’ which are not human actions (Anscombe,
undated a, pp. 208–209; 1989, pp. 127–138). I will call these
involuntary movements. In Intention Anscombe gives the following
examples: ‘The odd sort of jerk or jump that one’s whole body
sometimes gives when one is falling asleep’; ‘tics, reflex kicks from
the knee, the lift of the arm from one’s side after one has leaned
heavily with it up against the wall’ (§7, p. 13).

She defines this class as follows:

The class of movements of the body, in a purely physical description,
which are known without observation, and where there is no such
thing as a cause known without observation.

(§8, p. 15)

Just what Anscombe means by saying that in cases of involuntary
movement there is ‘no such thing as a cause known without obser-
vation’ is best seen by looking at the contrast class of involuntary
actions. This is because Anscombe distinguishes involuntary actions
as those in which there is such a cause (that is, a cause known
without observation).

To the class of involuntary actions, belong examples like the
problem case in §5: ‘Why did you knock the cup off the table?’
answered by ‘I thought I saw a face at the window and it made me
jump’. Other examples Anscombe gives are: ‘Why did you jump?’,
‘The leap and bark of the crocodile made me’ (§8, p. 15) and ‘He
withdrew his hand in a moment of recoil’ (§7, p. 13) (e.g ‘Why did
you withdraw your hand?’, ‘Suddenly the thought of touching him
repelled me’). Anscombe says that we have difficulty in describing
the distinction between a cause and a reason when ‘the cause itself
qua cause (or perhaps one should say the causation itself ) is in the
class of things known without observation’ (§9, p. 16).8

INTENTIONAL ACTION94

Anscombe introduces the label ‘mental cause’ for the kind of
causal statement she wants to isolate as causation known without
observation’. She says:

A ‘mental cause’ is what someone would describe if he were asked
the specific question: what produced this action or thought or feeling
on your part: what did you see or hear or feel, or what ideas or images
cropped up in your mind and led up to it?

(§11, pp. 16–17)9

It is clear enough that each of the examples we have given involve
a description which fits this criterion. ‘I saw a face at the window’,
‘The leap and bark of the crocodile (which I saw and heard)’, ‘The
thought “How repulsive!”’: each of these describe what was seen,
heard, felt, or thought of. However, what is not clear is with what
right Anscombe calls such statements causal.

There are two things about the idea of a ‘mental cause’, as defined
by Anscombe, that one might immediately object to. First, that
there is a kind of causation of human action that can be ‘known
without observation’; second that a statement like ‘I jumped
because I saw a face at the window’ is casual.

I will take these objections in reverse order because what I say
about the second will go some way toward addressing the first.

Anscombe herself notes that her talk of ‘causality’ in this context
is radically non-Humean.

This sort of cause of a feeling or reaction may be reported by the person
himself, as well as recognised by someone else … Note that this sort
of causality or sense of ‘causality’ is so far from accommodating itself
to Hume’s explanations that people who believe that Hume pretty
well dealt with the topic of causality would entirely leave it out of their
calculations; if their attention was drawn to it they might insist that the
word ‘cause’ was inappropriate or was quite equivocal. Or conceivably
they might try to give a Humian [sic] account of the matter as far as
concerned the outside observer’s recognition of the cause; but hardly
for the patient.

(§10, p. 16)

INTENTIONAL ACTION 95

Again, the concept of ‘cause’ requires its own, distinct, investiga-
tion, and Anscombe did write extensively on the topic (e.g. 1969b,
1971, 1974c, 1974d, 1977). Again, those enquiries resist easy
summary and require precisely the kind of painstaking attention
Intention demands. However, we can take from Anscombe’s ‘The
Causation of Action’ (1983a) some hints that will go some way
toward explaining the move Anscombe is making here.

Anscombe, in that paper, describes two ‘distinct directions of
enquiry’ that we might follow if we ask ‘How?’ or ‘Why?’ about
a human action, e.g. ‘Shutting the door’, where this question is
concerned with the cause of the action. In that paper she explicitly
excludes ‘explanation of action by intention’ saying that it ‘does
not properly come under my title, “The Causation of Action”’
(Anscombe, 1983a, p. 107). We should keep this point in mind.

The first direction of enquiry, she says, is one in which ‘we are
interested in picking out “chains” of causality going back in time’
(Anscombe, 1983a, p. 92). This investigation takes us in and in;
inside the physiological workings of a human body:

[The door shut] because of the push of a hand; that, because of the
placing of the hand and the extension of the arm; that, because of the
contraction of the muscles; this last because of the message down
the efferent nerves; this last … because of the afferent nerve impulses
leading to the sensory cortex and other parts of the brain ….

(Anscombe, 1983a, pp. 92–93)

The second direction of enquiry involves a ‘step back’ (Anscombe,
1983a, p. 97); it takes us out and out to the wider context of the
movement and requires ‘[r]ecognizing that is recognizing a pat-
tern of a different sort from that of elementary causation’. For
example: recognising that ‘the action was an immediate reaction
to an external stimulus; was, e.g., obeying an order’ (Anscombe,
1983a, p. 93).

The relation between these kinds of explanation is, of course,
a point of fierce philosophical dispute. Anscombe argues, in that
paper, that the first kind of causal explanation – in terms of phys-
iological processes – is dependent on our being able to give the
second kind in terms of what we can now call ‘mental causation’.

INTENTIONAL ACTION96

This is necessary, she says, ‘to know [what] is relevant’ (Anscombe,
1983a, p. 93). For example, if I am seeking to explain the door
shutting, I need to know the mental cause – Smith’s command
‘Shut that door!’ – in order ‘to know that the impulses in the
afferent nerves are relevant’ (Anscombe, 1983a, p. 93). Anscombe
compares this procedure to recognising the figure in a pointil-
liste picture in order to join up a line of dots (Anscombe, 1983a,
p. 93). Suppose I tell you ‘Trace a line of dots around the mon-
key in Seurat’s “A Sunday Afternoon on the Island of La Grande
Jatte”’. You will only be able to do so if you ‘step back’ every now
and again and see that figure. The figure tells you where to look
for your next dot.

A Humean will insist that this point is ‘methodological’; this is
to insist that an explanation in terms of – what Anscombe calls –
mental causality is not a genuine causal explanation. Anscombe
puts this objection in the mouth of her interlocutor: ‘when
we’ve got the information we want, we shan’t need to step back’
(Anscombe, 1983a, p. 97). The interlocutor insists that once we
know enough about the first kind of causal explanation, there will
be no need for the second. In Chapter 6 we return to the ques-
tion of whether explanation by mental causation and, indeed, by
intention, is more than ‘a highly convenient, nay indispensable,
facon de parler’ (Anscombe, 1983a, p. 103).

With this framing in mind, we can now say: the class of invol-
untary movements – reflex kicks, tics, and so forth – are those for
which there are only causal explanations in terms of ‘elementary
causation’. These explanations will always be known only by obser-
vation. And, we can now see that the objection that causation of
action cannot be known ‘without observation’, and the objection
that ‘Because of the bark of the crocodile’ is not really a causal
explanation, come from the same place. Both are a denial that
there are causal explanations ‘for the patient’ (§10, p. 16), i.e. ones
that irreducibly relate to the agent’s own account of what it was
that she was responding to – an order, a thought, an itch, a flash of
anger – when she acted as she did. The claim that ‘we shan’t need
to step back’ is the claim that causal explanation requires obser-
vation of physiological process. The claim that ‘stepping back’
is a necessary part of action explanation insists that how things

INTENTIONAL ACTION 97

were from the perspective of the agent is an essential part of the
causal-explanatory process.

It would take us too far from the topic of Intention to defend
Anscombe’s use of the label ‘causation’ for the kind of explana-
tion she has isolated. We can sidestep this problem by focusing on
the fact that Anscombe is seeking to describe a case in which we
speak of causation, and in which it is unclear what difference we
are marking when we do so. If she can explain what we mean by
speaking of causation in those contexts, then she has described a
kind of causation that we recognise. A debate about whether that
is really causation, belongs to a quite different project and, I think
for Anscombe, one would need to be very clear about what one
meant by ‘really’.

INTENTION, MOTIVE, CAUSE,
REASON (§§10–14)

We are now at the end of the first stage of Job A: the question
‘Why?’ lacks application when the answer gives a cause. We
are now entitled to use the notion of ‘cause’ without begging
any questions. As we have seen from the task of describing that
notion, it really could not be left intuitively obvious. Anscombe
has described without circularity two kinds of answer to the ques-
tion ‘Why?’ which mention past history, but which give a cause
rather than a reason. These answers classify the description as of
either an involuntary movement or an involuntary action.

A clear account of mental cause, one which does not rely on
metaphor, or our ‘intuitive’ understanding, is an essential pre-
liminary for approaching Job B, in which Anscombe aims to
‘explain the sense’ of the question ‘Why?’. For Anscombe, the
point of discussing mental causation is wholly negative: she has
described when an answer gives a mental cause simply in order
to distinguish such answers from those in which she is interested;
that is, answers which give a reason, motive or intention, and
hence show the question ‘Why?’ to have application (§11, p. 18).
This is one reason why she says, in §11, that having ‘isolated
[a] notion of mental cause’ she ‘believe[s] it is of very little
[importance]’ (p. 18).10

INTENTIONAL ACTION98

§§10–14 introduce three kinds of answer to the question
‘Why?’ which give a reason for acting: ‘the answer may (a) simply
mention past history, (b) give an interpretation of the action, or
(c) mention something future’ (§16, p. 24).

The choreography of these passages is extremely difficult. This
is because we have three interconnected notions in play: reason for
acting, motive, intention. Anscombe’s first concern is to distinguish
each of these from mental cause. A large part of her discussion aims
to show that though a statement giving a motive or intention or rea-
son can often be expressed using a sentence which could also be used
to make a statement of mental causality, this does not show that
these notions are equivalent. The danger of assuming such equiv-
alence is especially present when we speak about motives, because:

a very natural conception of ‘motive’ is it is what moves (the very
word suggests that) – glossed as ‘what causes’ a man’s actions etc’.
And ‘what causes’ them is perhaps then thought of as an event that
brings the effect about – though how it does – i.e. whether it should be
thought of as a kind of pushing in another medium, or in some other
way – is of course completely obscure.

(§11, p. 18)

But at the same time as applying the reason/cause distinction,
Anscombe also describes the relations between reason for acting,
motive and intention. These relations are complex. As Anscombe
says, sometimes ‘[a]sked for a motive, a man might say … “I did it
in order to …”’ which is just to give his intention in acting (§12,
p. 18); however, ‘there is an application of “motive” other than the
application of “the intention with which a man acts”’ (§12, p. 20).

To make matters more complex still, Anscombe identifies
three kinds of motive: backward-looking, motive-in-general,
and forward-looking. It is tempting to map these three kinds
of motive directly onto the three answers – (a) simply mention
past history, (b) give an interpretation of the action, or (c) men-
tion something future – but this would be a mistake. It is true
that answers falling under (a), (b) and (c) may show that what is
being done is being done for a backward-looking, interpretative,
and forward-looking motive respectively, but this implies neither

INTENTIONAL ACTION 99

that those answers give such motives nor that all such answers
are motive-giving. Sometimes they give a reason or an intention
but not a motive. Once again, we are moving from simplicity to
complexity.

I will take as our thread through these passages the three kinds of
reason-giving answers – (a), (b) and (c) from the summary in §16.

(A) SIMPLE MENTION OF PAST HISTORY

Recall how, in §5, Anscombe highlighted the difficulty of giving
a non-circular account of the distinction between two kinds of
answer falling under (a): ‘I thought I saw a face at the window
and it made me jump’ as an answer to ‘Why did you knock the
cup off the table?’ and ‘He killed my father’ as an answer to the
question ‘Why did you kill him?’. Having given her account of
mental causation, Anscombe has a much clearer sense of the dis-
tinction she needs to make and can now explain the source of the
difficulty. When I describe a past event in giving a reason for my
action this is also a statement which belongs to the class of things
known without observation. So, now the task is to differentiate
two classes of answers to the question ‘Why?’, both of which are
(1) descriptions of past events and (2) not made on the basis of
observation. Put this way, the distinction between a mental cause
and this kind of reason for acting looks much more slight than we
might have supposed (if, for example, we had assumed a notion of
‘cause’ which would satisfy a Humean). As we will see, Anscombe
thinks that sometimes there is no distinction at all.

There is a kind of progress here that is similar to that marked
in §2 (p. 2). There Anscombe’s enquiry had begun by seeking a
distinction between prediction and expression of intention, but
by giving a criterion for the former, it turned into something else:
‘The “intuitively clear” distinction we spoke of turns out to be
a distinction within the class of predictions between expressions
of intention and estimates’. So too here: an ‘intuitively clear’ dis-
tinction between reasons and causes turns out to be a distinction
within the class of answers which are descriptions of past history
known without observation.

INTENTIONAL ACTION100

Anscombe gives the following account of the distinction:

If an action has to be thought of by an agent as doing good or harm
of some sort, and the thing in the past as good or bad, in order for
the thing in the past to be the reason for the action, then this reason
shews not a mental cause but a motive. This will come out in the
agent’s elaborations on his answer to the question ‘Why?’

(§14, p. 22)

By this criterion, ‘He killed my father’ ‘shews’ – though, as we
will see, does not give – a motive, while ‘He jumped out of the
broom cupboard’ gives a mental cause. One way to test whether
this connection with good and harm holds in a particular case,
says Anscombe, is to see whether the explanation is given up or
disavowed ‘if you could e.g. show that either the action for which
he has revenged himself, or that in which he has revenged himself
was quite harmless or was beneficial’ (§14, p. 21–22). We might
imagine someone saying: ‘I didn’t invite Ann to the party because
she ignored my letter’, but then acknowledging that he had had
no reason not to invite Ann once it is pointed out to him either
that the letter had never reached her or that Ann despised party
invitations. If ‘he ceases to offer a reason’ in such a case, or says
‘I thought that was a reason’ this shows that we are dealing with
motive not a cause as ‘no such discovery would affect an assertion
of mental causality’ (§14, p. 22). Pointing out that there is no
reason to jump at sudden noises does nothing to undermine the
causal statement, ‘I jumped because of the loud bang’.

Above I pointed out that while ‘He killed my father’ ‘shews’ a
motive, it does not give a motive. This distinction is important
in helping to keep clear the difference between motive and cause.
‘He killed my father’ gives my reason, and at the same time shows
that I act for the backward-looking motive revenge. The motive
of revenge looks backward to some harm that was done to me,
which I now take as a reason to harm; that I kill because he killed
my father characterises my murder as revenge. This means that an
action can be characterisable as revenge even when the agent does
not have the concept of revenge; if he harms someone because
she harmed him, then this is the ‘primitive, spontaneous’ case

INTENTIONAL ACTION 101

(§35, p. 66), and he can be described as taking revenge. A more
sophisticated form of revenge involves taking revenge as one’s
object. As when one’s intention is to get my revenge (§35, p. 66).
In such a case one recognises the harm one was done as a reason
for getting revenge, rather than as merely a reason for doing harm.

Anscombe says that when a reason is given up in the light of
evidence that, e.g. ‘the action for which he has revenged himself …
was quite harmless or was beneficial’ this also identifies a third
class of involuntary actions: ‘If I do something out of revenge
which is in fact advantageous rather than harmful to my enemy,
my action, in its description of being advantageous to him, is
involuntary’ (§14, p. 22).11 It is clear that this kind of involun-
tary action could not be described without using the concept of
intention – given its dependence of the notion of a reason for
action, in the form of a backward-looking motive. This fact vali-
dates Anscombe’s earlier insistence that it would be illegitimate to
‘introduce “It was involuntary” as a form for rejecting “Why?” …
when the whole purpose of the elucidation was to give an account
of the concept “intentional”’ (§7, p. 13). It is also clear that an
action that was involuntary in this sense would not be one for
which the agent could take credit (if they did good in seeking
to harm), though they might be exculpated (if they did harm in
seeking to do good).12

Having made the distinction between a description of a past
event which gives a reason and one that gives a cause Anscombe
says – and as a reader it is hard not to be dismayed by this – that
though it exists, and in many cases it is clear, it is also true that in
‘many cases the distinction would have no point’ (§15, p. 24) and
‘that reason and cause are everywhere sharply distinct notions, is not
true’ (§15, p. 24). Note that if this is right, then the idea that the
class of intentional actions can be sharply delineated must also be
given up. She introduces this thought with the following example:

Why did you do it?
Because he told me to.
Is this a cause or a reason? It depends very much on what the action
was or what the circumstances were.

(§15, p. 23)

INTENTIONAL ACTION102

Relevant factors would include the test we have just given – whether
the explanation is given up when it is explained, for example, that
‘he’ is a bad person and shouldn’t be relied upon for practical guid-
ance. But also, ‘how sudden one’s reaction was’, whether the agent
thought of it as ‘having a significance’ (§15, p. 24). For example, if
one acted out of pity, knowing that no one ever did what he said.

The fact that Anscombe’s distinction entitles her to acknowl-
edge an area in which the distinction between cause and reason
has no point or cannot be made, and to explain why such cases
exist, illustrates how unamenable to a Humean treatment is
Anscombe’s notion of ‘mental causation’. The concept of mental
causality, as we saw, is connected with the personal; it is ‘known
without observation’ and is distinguished sharply from physiolog-
ical explanation or explanation in terms of efficient causes. This
means that there is nothing that an empirical – i.e. scientific –
investigation might turn up that would be relevant to the question
‘Reason or cause?’.

And the account Anscombe has given can also explain why
these cases exist. Because a mental cause can be given by an answer
which mentions what you ‘see or hear or feel, or what ideas or
images cropped up in your mind’ and because these are things that
may or may not have significance for an agent, or which are such
that their significance (or whether they have it) may be opaque,
unclear or ambiguous, the distinction between reason and cause
may be unclear even to oneself.

The class of backward-looking motives plays an interesting role
in Intention. There are two points at which Anscombe mentions
the importance for us of the practice we have just described –
that of discussing, criticising or endorsing, the motives that a
person has (§14, p. 22; §20, p. 31). In both cases she highlights
in particular their importance in relation to someone’s proposed
actions. So far we have only looked at current or past actions. One
thought might be that our interest in a man’s backward-looking
motives is especially tied to our practices of seeking to influence
each other’s behaviour. There is not much point in criticising a
man’s motives if he has already acted, but if he is proposing to act
then getting him to think differently about the thing in the past
(i.e. that it was good rather than bad, or vice versa) or getting

INTENTIONAL ACTION 103

him to think differently about the action he proposes (that it
will do harm rather than good, or vice versa) is a way of getting
him to give up his reasons. We will consider this issue again in
Chapter 7.

(B) INTERPRETATION OF THE ACTION

Anscombe says, ‘To give a motive [in general] is to say something
like “See the actions in this light”’ (§13, p. 210). For example, ‘I
did it out of love’ or ‘See it as an act of friendship’. Anscombe says
that such answers are ‘expressive of the spirit in which [something
was done] rather than descriptive of the end to which the [act]
was a means’ (§12, p. 18). These motives, she seems to be saying,
do not give the intention with which a person acts; ‘I did it out
of revenge’ and ‘I did it out of friendship’ have, on Anscombe’s
view, very different logical or grammatical forms. The former is
equivalent to ‘I did it in order to be revenged’ but the latter cannot
be rendered in the ‘in order to’ form. That does not mean that
one can’t make ‘getting him to be my friend’ or ‘making a public
display of friendship’ as one’s object, but if one did friendship
would not be a ‘motive in general’ but rather securing or display-
ing friendship would be one’s end.

Anscombe says very little about interpretative motives, sug-
gesting that the topic belongs to ‘ethics or literary criticism’
(§12, p. 19) Having invited a reader to ‘Consider the statement
that one motive for my signing a petition was admiration for
its promoter, X’ – a statement which gives an ‘interpretative
motive’ – Anscombe almost straightway interrupts her discus-
sion: ‘I say “Consider this” really with a view to saying “let us
not consider it here”. It is too complicated’ (§13, p. 20). She
doesn’t say why it is too complicated, nor why it’s being compli-
cated means the discussion can be abandoned. One suggestion
I have is that these answers are ‘too complicated’ because they
involve one taking an interpretative stance toward one’s own
intentional actions, and as such involve a kind of complexity
that goes beyond – or at least takes us beyond the core of – the
linguistic practice Anscombe is seeking to describe. This is per-
haps why they belong to ‘literary criticism’ (§12, p. 19). When

INTENTIONAL ACTION104

Wittgenstein speaks of these interpretative motives he connects
them with a complex account of ‘the whole history of the incident’:

I want to tell him something about myself, which goes beyond what
happened at that time. I reveal to him something of myself.

(Wittgenstein 1958, §659; §25, p. 45)

Such explanations are clearly important in our interpersonal rela-
tions – how many marital arguments turn on a dispute about how
what one did or said should be interpreted? – but are not central
so far as the topic of intention goes. One way to see this is to
recognise the unintelligibility of a person whose only answer to
a question ‘Why did you kill him?’ was an interpretative motive.
Such explanations, I think, must go alongside, or be given in the
light of, reasons falling into the other categories.

(C) MENTION OF SOMETHING FUTURE

We know that Anscombe thought it important to define the class
of answers giving mental causes in order to ‘distinguish it from the
ordinary senses of “motive” and “intention”’ (§11, p. 18). In the
intuitive voice, she describes the difficulty:

Now one might think that when the question ‘Why?’ is answered by
giving the intention with which a person acts – for example by men-
tioning something future – this is also a case of a mental cause. For
couldn’t it be recast in the form: ‘Because I wanted …’ or ‘Out of a
desire that …’?

(§11, p. 17)

This is the voice of someone who wants to insist that the relation
between reason and action is a causal-psychological one.

Anscombe can now say: not every case of ‘Because I wanted …’ is
an answer to the question ‘what produced this action or thought or
feeling on your part: what did you see or hear or feel?’ (§11, p. 17).
Sometimes ‘I wanted to …’ describes a feeling – in which case it
gives a mental cause – but sometimes ‘I wanted to’ introduces the
object of one’s action. When ‘I wanted to’ introduces the object of

INTENTIONAL ACTION 105

one’s action, this has nothing to do with describing a feeling. The
difference between these two species of wanting will be crucial
later, when we come to the topic of practical reason.

To clarify the difference, Anscombe describes her third class
of motives, ‘forward-looking motives’. She says: ‘I call a motive
forward looking if it is an intention’ (§13, p. 21), which is not, of
course, to say that all intentions are forward-looking motives. She
introduces the notion with an example:

For example, to say that someone did something for fear of … often
comes to the same as saying he did so lest … or in order that … should
not happen.

(§13, p. 21)

This example is at first rather surprising. When someone does
something because they are afraid, this is most naturally thought
of as a reaction to a fearful object, so it seems then badly suited
to the purpose of displaying a kind of motive that is an intention.
‘I was afraid of him’ sounds more like an answer giving a mental
cause than an intention.

However, Anscombe uses the example of fear precisely to illus-
trate the distinction between the cause of an action and the object
of an action, a distinction which parallels that between a cause of
fear and an object of fear:

A child saw a bit of red stuff on a turn in the stairway and asked what
it was. He thought his nurse told him it was a bit of Satan and felt a
dreadful fear of it. (No doubt she said it was a bit of satin.) What he
was frightened of was the bit of stuff; the cause of his fright was the
nurse’s remark. The object of fear may be the cause of fear, but, as
Wittgenstein remarks,13 is not as such the cause of fear.

(§10, p. 16)

Asked ‘Why did you scream?’, my answer ‘I felt afraid’ would give
a mental cause; asked ‘Why did you stay up all night revising?’,
‘I was afraid I would fail the exam’ gives the object of my staying
up: ‘In order to not fail the exam’. The first describes a feeling the
second introduces an object of desire.

INTENTIONAL ACTION106

In §16 Anscombe ‘summariz[es] conclusions reached so far’ (p. 24).
The summary gives a sense of the ‘chaos’ created through her
proposed ‘clarification’ of the notion of ‘reason for acting’.

Intentional actions are a subclass in a man’s history which are known
to him not just because he observes them. In this wider class is
included one type of involuntary actions, which is marked off by the
fact that mental causality is excluded from it; and mental causality is
itself characterised by being known without observation. But inten-
tional actions are not marked off just by being subject to mental
causality, since there are involuntary actions from which mental cau-
sality is not excluded. Intentional actions, then, are ones to which the
question ‘Why?’ is given application, in a special sense so far explained
as follows: the question has not that sense if the answer is evidence
or states a cause, including a mental cause; positively, the answer may
(a) simply mention past history, (b) give an interpretation of the action
or (c) mention something future.

(p. 24)

It is hard not to panic at this point. However, we must be patient.
As we will see, when Anscombe turns to her second head, ‘inten-
tion with which’, an order will emerge in the class of descriptions
she has here described.

VOLUNTARY VS INTENTIONAL
ACTIONS (§§17–19)

Before that, however, Anscombe has one final addition to make.
She says she will ‘complete [her] account of when our question
“Why?” is shewn not to apply’ (§17, p. 25). This is the second
part of task [A]; here Anscombe identifies answers to the question
‘Why?’ which are known without observation but which give nei-
ther a mental cause nor a reason for acting. These two paragraphs
seem rather isolated, a sort of coda to the main work undertaken
between §5 and §16. For the sake of simplicity and summary it
is tempting to neglect them, especially because what emerges in
these two sections is a rather awkward and unwelcome amendment
to the description with which Anscombe began her enquiry into

INTENTIONAL ACTION 107

intentional action. Recall that she said that intentional actions are
‘actions to which a certain sense of the question “Why?” is given
application; the sense is of course that in which the answer, if pos-
itive, gives a reason for acting’ (§5, p. 9). Her job up to this point
has been clarifying the range of answers which give reasons for
acting and hence show the question has application. Now she says:

the answers to the question ‘Why?’ which give it application are …
more extensive in range than the answers which can be given as
reasons for acting.

(§18, p. 28)

Here is another clear point at which Anscombe is prepared to
sacrifice narrative arc and reader comfort rather than leave ends
untied or cases undifferentiated.

Tempting as it is, for the reason just highlighted, to draw a dis-
creet veil over this pair of paragraphs, the distinctions they contain
are important. In them Anscombe describes, albeit provisionally,
one class of actions that are ‘voluntary, rather than intentional’
(§17, p. 26); as we saw earlier, this class is of crucial importance
to Catholic moral psychology and to Anscombe’s wider ethical
project. She also reflects on the fact that declining to give a rea-
son for action is ‘often quite intelligible, sometimes strange; and
sometimes unintelligible’ (§18, p. 26). Both these points have a
role to play later in Intention.

Anscombe notes that not every intelligible answer to the ques-
tion ‘Why?’ gives a reason for acting.

Now of course a possible answer to the question ‘Why?’ ‘I just thought
I would’ or ‘It was an impulse’ or ‘For no particular reason’ or ‘It was
an idle action – I was just doodling’. I do not call an answer of this sort
a rejection of the question. The question is not refused application
because the answer to it says there is no reason, any more than the
question how much money I have in my pocket is refused application
by the answer ‘None’.

(§17, p. 25)

The question ‘How much money do you have in your pocket?’
is refused application by the answer ‘I don’t have any pockets’;

INTENTIONAL ACTION108

the question presupposes my having pockets but not my having
money in them. So Anscombe’s suggestion seems to be something
like: the question ‘Why?’ does not presuppose a reason for acting,
but only that what is happening is the sort of thing for which a
reason might be given. Those answers which reject the question
‘Why?’ reject this presupposition – if what is happening is known
to me only by observation or if I only know its cause by observa-
tion, then it is not the sort of thing for which a reason might be
given – but ‘No reason’ does not.

Anscombe says that answers in this range are a ‘curious inter-
mediary case’:

The question ‘Why?’ has and yet has not application; it has application
in the sense that it is admitted as an appropriate question; it lacks it in
the sense that the answer is that there is no answer.

(§17, p. 26)

She notes an associated answer, ‘I don’t know why I did it’, which
she says is ‘appropriate to actions in which some special reason
seems to be demanded, and one has none’ (§18, p. 25).

Anscombe says that answers in this range are ‘often quite intel-
ligible; sometimes strange; and sometimes unintelligible’ (§18,
p. 26). My suggestion in Chapter 3 was that Davidson’s would-be
squirrel house builder might fail to be intelligible, or at least be
strange, in just this way. Intention gives an example of such unin-
telligibility:

[I]f someone hunted out all the green books in his house and spread
them out carefully on the roof, and gave one of these answers to the
question ‘Why?’ his words would be unintelligible unless as joking or
mystification. They would be unintelligible, not because one did not
know what they meant, but because one could not make out what the
man meant by saying them here.

(§18, p. 27)

It is clear enough that Anscombe is right about this – his words
would be unintelligible. But why? Someone who explained his

INTENTIONAL ACTION 109

going along with a rather unpleasant plan to humiliate a colleague
might say ‘I don’t know why I did it’ and be quite intelligible.
What is the difference?

It is tempting at this point to think that we are coming up
against a psychological fact, but we should recall Cora Diamond’s
remark:

The capacities picked out by ‘the mind’ are those for coherence and
commitment in thought and language, they are capacities characteris-
tically exercised in going on intelligently and intelligibly.

(Diamond, 1991, p. 5)

As Intention progresses we will come to understand that the place
to locate the difference is in the action description and in facts
that belong to the ‘natural history of mankind’, rather than in
anything that is going on at the time of the action. We can begin
to see this if we note that ‘hunting out all the green books in his
house and spreading them out carefully on the roof ’ is not the sort
of behaviour that is normal, that has a point that connects with
the ordinary sorts of things a man might want or enjoy, that we
might naturally go in for. As such, it seems to demand a particular
explanation. To make sense of it we want to know what it was
for, why he wanted to do it. In contrast, going along with a nasty
plan one comes to regret is the sort of behaviour we recognise as
quite ordinary, if saddening. In the next chapter we will see how
Anscombe starts to formalise the intelligibility conditions that
surround the practice of giving reasons for action.

NOTES

1 Hursthouse offers the following ‘general version of the causalist thesis:
‘Intentional actions are actions (or movements) caused, perhaps in
a certain way, by certain mental states or events, whose occurrence
explains the occurrence of the action or movement’ (Hursthouse,
2000, p. 84). This thesis is part of the predominant view in the philos-
ophy of action.

2 Here is another point at which we should read Anscombe in light of
Aquinas.

INTENTIONAL ACTION110

3 See also, Anscombe (1989), esp. pp. 127–138. The distinction, along
with the beard-stroking example, is Aquinas’ in Summa Theologiae
I–II, I, I).

4 See Hamilton (2013).
5 See Cora Diamond (1966, p. 232), for a discussion of ‘internal

description[s] … that use […] a word taken from elsewhere’ and other
metaphorical uses of language.

6 See Schwenkler (2015) for a discussion of this point in the context of
Anscombe’s Thomism.

7 Compare Anscombe, 1964, pp. 39–40.
8 I take it that this parenthesis is to make clear that it is the causality and

not the causal antecedent that is known without observation. ‘There
was a face at the window’ is clearly an observation statement while, if
Anscombe is right, ‘I jumped because there was a face at the window’
is not.

9 As should be clear, a ‘mental cause’ could be a non-mental event, e.g. a
knock at the door or a mental event, e.g. a sudden thought. If the former,
however, ‘it must be something perceived by the person affected –
e.g. the knock on the door must be heard – so if in this sense anyone
wishes to say it is always a mental event I have no objection’ (§11, §17).

10 Another is that when a request for explanation of an action is answered
by giving a mental cause this, Anscombe thinks, leaves it entirely open
whether the action, under that description, is intentional. Recall her
remark in ‘Causation of Action’ that ‘explanation of action by intention’
does ‘not properly come under my title’. Such an answer is not an
answer to the question ‘Why are you X-ing?’ or ‘Why did you X?’ but
nor does it rule out such an answer in the way that ‘I didn’t know I was
X-ing’ would. For example suppose someone asks ‘Why did you leap
up?’. Your answer ‘I had a sudden thought “I must act now, before it
is too late”’ gives a mental cause, by Anscombe’s definition. But this
does not show that what was done was not intentional. Such an answer
would probably elicit a follow up enquiry: ‘Yes, but why did you leap up?
What for?’ Equally, if you are opening the window and I ask you ‘What
produced this action or thought or feeling on your part: what did you
see or hear or feel, or what ideas or images cropped up in your mind
and led up to it?’ you may well answer ‘I don’t know that there was any
definite history of the kind you mean’ (§11, p. 17) or ‘It merely occurred
to me …’. This claim that there was no mental cause – a claim which,

INTENTIONAL ACTION 111

note, by Anscombe’s lights leaves it entirely open that a story in terms
of ‘elementary causation’ could be told – does not show that my open-
ing the windows is not an intentional action; nor that it is. This is the
reason that mental causation is not a topic for Intention just as inten-
tion was not a topic for causation of action.

The idea that these different kinds of explanation are independent of
one another – that ‘I thought “That unspeakable man”’ (causal expla-
nation) and ‘In order to tidy up’ (reason for action) do not exclude one
another is potentially enormously illuminating. Anscombe does not,
though it would be fascinating to do so, map out the ways in which such
answers might impose limits of intelligibility on one another. For example,
could the two answers above be simulteously accepted when I ask you
‘Why did you burn his manuscript?’ Whether a similarly pluralist move
could be made in relation to kinds of causal explanation, e.g. ‘I thought
“That unspeakable man”’ (causal explanation) and ‘I was in physiological
state XYZ’ (physiological explanation) is also clearly of interest.

11 This is the fourth of Anscombe’s examples of the involuntary: ‘the invol-
untary benefit I did him by a stroke I meant to harm him’ (§7, p. 13).

12 This asymmetry between good and bad actions is a major strand in
Anscombe’s ethical writings.

13 Wittgenstein, Philosophical Investigations, §476. This reference is
Anscombe’s own.

SUGGESTED READING

Anscombe, G. E.M. (1962b). ‘On Sensations of Position’, Analysis, 22(3). Reprinted
in MPM, pp. 71–74.

—— (1965b). ‘The Intentionality of Sensation: A Grammatical Feature’. From
R. J. Butler (ed.), Analytical Philosophy, second series (Oxford). Reprinted in
MPM, pp. 3–20.

—— (1983a). ‘The Causation of Action’. From C. Ginet (ed.), Knowledge and Mind
(New York and Oxford: Oxford University Press), pp. 174–190. Reprinted in
HLAE, pp. 89–108.

Davidson, Donald (1963). ‘Actions, Reasons, and Causes’. In his Essays on Actions
and Events (Oxford: Oxford University Press, 2001), pp. 3–20.

Hursthouse, Rosalind (2000). ‘Intention’. In Roger Teichmann (ed.), Logic, Cause,
and Action (Cambridge: Cambridge University Press), pp. 83–106.

McDowell, John (2011). ‘Anscombe on Bodily Self-Knowledge’. In Anton Ford,
Jennifer Hornsby, and Frederick Stoutland (eds), Essays on Anscombe’s Intention
(Cambridge, Mass.: Harvard University Press), pp. 128–146.

INTENTIONAL ACTION112

Martin, C. B. (1971). ‘Knowledge without Observation’, Canadian Journal of Philosophy,
1(1; September), pp. 15–24.

Teichmann, Roger (2014b). ‘The Voluntary and the Involuntary: Themes from
Anscombe’, American Catholic Philosophical Quarterly, 88(3), pp. 465–486.

Vesey, G. N. A. (1963). ‘Knowledge Without Observation’, The Philosophical Review,
Vol. 72, pp. 198–212.

5
INTENTION WITH WHICH

(§§20–40)

In this chapter we follow Anscombe’s enquiry into intention with
which. In these sections the apparently haphazard set of distinc-
tions and classifications Anscombe set out in §§5–19 begins to
resolve into something coherent. Where the previous chapter saw
us move from simplicity to complexity, here we move from com-
plexity to order. The order, we will see, is one that could only
come into view once that complexity had been exposed.

The core of this chapter sets out the ‘calculative order’ which
Anscombe says ‘is there whenever actions are done with intentions’
(§42, p. 80). This order is not an order in ‘actual mental pro-
cesses’ but is a formal order which holds between descriptions
belonging to the class outlined between §5 and§19; as Anscombe
says, if it ‘were supposed to describe actual mental processes, it
would in general be quite absurd’ (§42, p. 80). Anscombe artic-
ulates this order from two directions: first, by considering a case
of intentional action; second, by setting out the formal character
of practical reasoning. A major question for this chapter is why
Anscombe is repetitive in this way. As we will see, while the first

INTENTION WITH WHICH114

direction of approach allows her to describe the calculative order,
the second enables her to dislodge an intuitive picture which
renders that description unintelligible.

Before turning to the calculative order, we should discuss a pair
of bridging sections between Anscombe’s statement that she has
‘roughly outlined the area of intentional actions’ and her announce-
ment in §22 that she is turning to her topic under its second head:
intention with which. In these sections Anscombe poses a question:
what is the relation between the three uses of ‘intention’ identified
in §1? As we will see, this question, and Anscombe’s answer to it,
has nothing in common with Anscombe’s Question.

INTENTIONAL ACTIONS, FURTHER
INTENTIONS IN ACTING, INTENTIONS
FOR THE FUTURE (§§20–21)

In these sections Anscombe poses a question about the ‘three
heads’ of the concept of intention. The question is not: what
is the connection between the different kinds of phenomena
to which we apply the concept intention? That is, it is not
Anscombe’s Question. Rather the question is: is it essential to
the concept of intention that we use the word ‘intention’ in these
three ways, or could we recognise the concept of intention as
part of a linguistic practice which used the word ‘intention’ only
in relation to actions (§20, p. 30)?1 Looking at these sections
will remind us of points about method that were our concern
in Chapter 3. It will also be an opportunity to make sharp a
distinction in Anscombe’s book – a distinction that though
important is easily missed – between further intentions in acting,
and intention in a proposed action.

To give one’s further intention – to give the intention with
which one is acting – is to give the description under which one
acts, and this is often to give a wider description of what one is
doing. An expression of intention for the future is, in contrast,
a description of a proposed action (§14, p. 22). Compare ‘I am
going to build a squirrel house’ said by Jones as he saws a plank,
and the same said by Smith when asked whether he has any plans
for the summer break. Jones is already building a squirrel house

INTENTION WITH WHICH 115

(though no squirrel house has yet been built by him). For him, the
future-tense reflects the fact that his action is not yet completed;
but he could just as well have spoken using the present-tense ‘I
am building a squirrel house’. For Smith, in contrast, building a
squirrel house is only some possible future action of his. For him,
the future-tense is essential.2

With this distinction in mind, let us turn to §20. Anscombe
asks:

would intentional actions still have the characteristic ‘intentional’, if
there were no such thing as expressions of intention for the future,
or as further intentions in acting? i.e. is ‘intentional’ a characteristic
of those actions that have it, which is formally independent of those
other occurrences of the concept of intention?

(§20, p. 30)

Having posed this question Anscombe then says she will ‘test this’
by making ‘two rather curious suppositions’:

(a) suppose that ‘intention’ only occurred as it occurs in ‘intentional
action’, and (b) suppose that the only answer to the question ‘Why are
you X-ing?’, granted that the question is not refused application, were
‘I just am, that’s all’.

(§20, p. 30)

The argument of §20 is extremely difficult to bring into focus,
even by the standards of Intention. The main interpretative chal-
lenges are to work out the relationship that these two suppositions
stand in to each other and also to the hypothesis (or hypotheses)
that Anscombe wishes to test.

It can be natural to take Anscombe to be testing a single hypo-
thesis (that ‘intentional’ only occurs in intentional action) and
so to assume that the two suppositions combine to impose this
restriction. What can make this natural is the fact, just noted, that
a further intention can often be expressed in the form ‘I am going
to do X’. As in ‘Why are you boiling the kettle?’, ‘I am going to
make tea’. However, this is not the only way that a further inten-
tion can be given. As we have already seen, a description of one’s

INTENTION WITH WHICH116

motive – backward- or forward-looking – can also give the further
intention with which one acts, without thereby describing some-
thing future.

[I]f I kill a man as an act of revenge I may say I do it in order to be
revenged, or that revenge is my object; but revenge is not some further
thing obtained by killing him, it is rather that killing him is revenge.

(§13, p. 20)

In the same way that ‘killing him is revenge’, ‘killing him’ may
also be ‘showing my peers I’m no phoney’, ‘starting a war’, ‘passing
the time of day’. None of these describe a proposed action but are
rather ‘wider description[s] of what [I am] doing’ (§22, p. 35).

My suggestion is therefore that we take the comma after
‘intention for the future’ in Anscombe’s question to be indica-
tive of a list: there are two hypotheses to be tested, each relating
to one of the other two uses of ‘intention’. To speak of further
intentions in acting is, in Intention, to speak of a ‘calculative order
that is there whenever actions are done with intentions’ (§42,
p. 80) – this calculative order is our topic in this chapter. To
speak of ‘expressions of intention for the future’ is to speak of
a special kind of prediction which occurs in our language and
which is to be distinguished from estimate for the future – our
topic in Chapter 7. Of course, these topics are not unconnected,
but they are different.

Once we distinguish two hypotheses, we can see that for each
hypothesis there is a supposition. The first hypothesis is that there
were no such thing as the sort of prediction that we call an expres-
sion of intention for the future. This hypothesis will rule out any
answer to the question ‘Why?’ which is not a description of what
one is doing or a past-tense reason-giving answer. As such it will
allow some answers which show backward-looking motives, but
not, e.g., ‘I’m going to get my revenge’. It will also allow wider
descriptions of what one is doing, but not descriptions of actions
that ‘we could object that he is not yet doing Y but only doing
X with a view to doing Y, as when a man takes his gun down
with a view to shooting rabbits’ (§20, p. 31). This fits supposi-
tion (a): ‘suppose that “intention” only occurred as it occurs in

INTENTION WITH WHICH 117

“intentional action”’. The second hypothesis is that there were
no such thing as a further intention with which a person does
what he does. This latter hypothesis will rule out backward- and
forward- looking motives, as well as wider descriptions of what
one is doing and further intentions in doing what one is doing.
That is, as per supposition (b) – ‘suppose that the only answer to
the question “Why are you X-ing?”, granted that the question is
not refused application, were “I just am, that’s all”’ – it will rule
out all reason-giving answers to the question ‘Why?’.

Anscombe argues that these hypotheses have different con-
sequences: on the first hypothesis, we are left with ‘a very thin’
notion of intentional action; on the second we are left without a
concept of intention at all (§20, p. 32). Let us try to see why.

The first hypothesis was that ‘there were no such thing as
expressions of intention for the future’. Anscombe tests this by
supposing that ‘“intention” only occurs in “intentional action”’
(§20, p. 30).

Under this supposition a description of what is happening,
‘X-ing’, will be a description of an intentional action if an answer
to the question ‘Why are you X-ing?’ gives a backward-looking or
interpretative motive, or a wider description of what is happening.
So it looks as if we are still able to recognise the concept as applied
to action.

The question Anscombe then asks is whether this gives us
‘enough to constitute intentional actions as a special kind’ (§20,
p. 31) and answers:

It seems reasonable to say that if the only occurrence of intention were
as the intention of doing whatever one is doing, the notion of inten-
tional action would be a very thin one; it is not clear why it should be
marked off as a special class among all those of a man’s actions and
movements which are known to him by observation.

(§20, p. 32)

Anscombe’s point, bear in mind, is not about any individual; it
is not a claim for empirical psychology. We can, I think, imagine
someone whose life is so impoverished that, for some time at
least, she never has intentions for the future. Perhaps all her

INTENTION WITH WHICH118

actions are motivated by events in her past and she has no future
state of affairs that she wants to bring about. Or perhaps she is
‘living in the moment’ to such a degree that she walks, drinks,
eats, converses, not in order to bring anything about but for its
own sake.

Anscombe’s claim seems rather to be that the interest of the
concept, our interest in a person’s intentional actions, as distinct
from her voluntary or involuntary ones – would vanish, or at least
diminish, if there were no such thing as an expression of intention
for the future. In such a scenario people would never describe what
they proposed to do, and would give their motives and further
intentions only when they had already acted on them. The only
context in which you could ‘argue against motives’ would be to
‘criticise a man for having acted on such a motive’ (§20, p. 31). I
think Anscombe’s point is something like: without ‘You shouldn’t
do that because …’, ‘You shouldn’t have done that because …’ loses
its point. We will consider this suggestion again in Chapter 7
when we look more closely at what Anscombe says about expres-
sions of intention for the future (in §§2–4 and §§50–52).

What now of the second hypothesis, that there is no such
thing as further intentions in acting though there are expressions
of intention for the future? On this hypothesis, expressions of
intention for the future are permitted only so long as they are
descriptions of proposed future actions and not wider descrip-
tions of what one is doing. They can still occur, as in ‘One of
these days I am going to sit down and write a long letter’, but
not as answers to the question ‘Why?’ where they would count
as giving a further intention in acting (e.g. ‘I am going to make
some tea’ said by someone who is filling the kettle). In this case,
says Anscombe, it will ‘no longer be possible to differentiate
within the class of acts known without observation’ (§20, p. 32).
This in turn implies that ‘there would be no distinction between
such things as starts and gasps and, quite generally, voluntary
actions’ (§20, p. 32).

This is a strange but important piece of argument. The thought
must be something like the following. Suppose ‘raising my arm’
is a description of my intentional action but, ex hypothesi, there is
no such thing as a wider description of what I am doing in raising

INTENTION WITH WHICH 119

my arm. Any answer which describes something in the past can
only be a mental cause without the notion of a motive available.
‘He killed my father’ can only be a cause if the action cannot
be described as getting revenge. This means that we will be left
with a class of descriptions of bodily movements known without
observation where ‘the only answer to the question “Why?” is “I
just am”’. Thus, concludes Anscombe, ‘the occurrence of other
answers to the question “Why?” besides one like “I just did” is
essential to the existence of the concept of an intention or invol-
untary action’ (§20, p. 33).3

Anscombe’s view, then, is that the very idea of an intentional
action depends on the notion of a further intention in acting
(i.e. doing Y in doing X). Without it there would be no concept of
intentional action at all. And her suggestion – which we follow up
in Chapter 7 – seems to be that without expressions of intention
for the future, though we could distinguish the class of intentional
actions the point of doing so would be impaired.

The idea that we have three guises of intention to connect is
thus totally misguided as a reading of the project of Intention;
according to Anscombe we cannot even describe the application
of the concept of intention to actions or the significance of that
practice without coming to see that we must speak of expressions
of intention for the future and of the intention with which a
thing is done.

Anscombe connects the conclusion about the relation between
intentional action and further intentions in acting with the
Ancient and Medieval argument ‘that human beings must always
act with some end in view’ (§21, p. 33). This claim, she says,
would be wrong if taken to be a psychological claim about indi-
vidual humans – a man can ‘just do what he does, a great deal of
the time [without] a reason or purpose’ (§21, p. 34). However,
taken as a conceptual claim – in Fregean terms: a claim about
‘the mind, not minds’; in Wittgenstein’s terms: a claim about ‘the
look of human life containing rules’ – the argument mirrors that
in §§20–21:

[W]e can now see why some chain must at any rate begin. As we have
seen, this does not mean that an action cannot be called voluntary

INTENTION WITH WHICH120

or intentional unless the agent has an end in view; it means that the
concept of voluntary or intentional action would not exist, if the ques-
tion ‘Why?’, with answers that give reasons for acting, did not.

(§21, p. 34)

This ‘chain’ is the topic of the remainder of this chapter.

DOING X IN ORDER TO DO Y (§22)

When she was seeking to outline the area of intentional actions,
Anscombe said that ‘If the answer to the question “Why?” is a
simple mention of something future, then it expresses the inten-
tion’. As we know, there are many answers that express intentions
but do not mention something future. It is also the case that not
every answer to the question ‘Why?’ that mentions something
future expresses an intention (§22, p. 34). ‘I’m going to be sick’ as
an answer to ‘Why are you getting that bucket?’ rarely expresses
an intention to be sick; nor ‘I am having garlic for supper’ as an
answer to ‘Why are you buying breath-freshener?’

By looking at when an answer mentioning something future
counts as an expression of intention, Anscombe begins to describe
the form of the ‘chain’ that she thinks is essential to intentional
action. This is to reveal the order in the class of descriptions picked
out by the question ‘Why?’ Looking at the future tense helps to
bring out this form, but, as we will see, she finds it in current
action too. Here is an example from earlier:

I put ink on paper in the form of letters. I am writing something. I am
in fact signing something with my name. And I’m thereby joining in a
petition to the governor of the state – or prison – where I am important.
I am taking part in a campaign to get people tortured under interroga-
tion. In doing this I am keeping a promise. I am avoiding trouble with
some conspirators who have got me to promise to do that.

(Anscombe, 1993, p. 149)

Anscombe illustrates a case in which an answer to the question
‘Why?’ which describes something future does not express an
intention with the following example:

INTENTION WITH WHICH 121

‘Why are you setting up a camera on this pavement?’ ‘Because Marilyn
Monroe is going to pass by’. That is just a statement of something
future but by no means expresses that I am setting up a camera with
the intention that Marilyn Monroe shall pass by. On the other hand, if
you say ‘Why are you crossing the road?’ and I reply ‘I am going to look
in that shop window’, this expresses the intention with which I cross
the road. Now what is the difference?

(§22, p. 35)

She suggests the following criterion: ‘the future state of affairs
mentioned must be such that we can understand the agent’s think-
ing it will or may be brought about by the action about which he
is being questioned’ (§22, p. 35). Anscombe formalises this in a
‘vague and general formula’:

In order to make sense of ‘I do P with a view to Q’, we must see how
the future state of affairs is supposed to be a possible later stage in the
proceedings of which the action P is an earlier stage.

(§22, p. 36)

Note that the intelligibility Anscombe is talking about here is
that which she described in §18: it is not that the words as such
would be unintelligible – we can all understand the sentence
‘I am setting up a camera with a view to Marilyn Monroe pass-
ing by’ – but we would struggle to understand a man who used
them (§18, p. 27). His words – which is to say, he – would be
unintelligible.

In order to make sense of a man who says he is setting up
a camera with this end in view we would need to see how he
could think that his setting up a camera might bring it about that
Monroe passed by. This is not a psychological constraint, but a
logical one. Anscombe calls such an account ‘filling in’. For exam-
ple, if he explained that Monroe is so publicity-hungry that she
will seek out and pass by any camera within a ten-mile radius,
then we could understand his putting up the camera in order
that she should pass by. Often, of course, this filling out is not
required: to think that crossing the road is a stage in proceedings
whose terminus is looking in the shop window on the other side

INTENTION WITH WHICH122

of the street is quite clearly intelligible. In Chapter 6 we will look
more closely at why certain kinds of answer require ‘filling out’
while others do not.

There are two features of Anscombe’s formula that should be
highlighted. First, the hypothetical ‘If I do P then Q’ is not a
causal or counterfactual statement, nor a statement of the form:
always, if I do P then Q. Anscombe gives the example of ‘I am
going upstairs with a view to fetch my camera’. It is not as if going
upstairs causes my camera to be fetched, nor that if I didn’t go
upstairs my camera wouldn’t be fetched (I could send you up to
get it), nor ‘that going upstairs usually produces the fetching of
cameras’ (§22, p. 36). Nevertheless, we can often make sense of
someone thinking that ‘the future state of affairs Q [my camera
being fetched] is supposed to be a later stage in proceedings of
which the action P [going upstairs] is an earlier stage’ (§22, p. 36).
Again, just how it is that we can ‘make sense’ of this is a question
Anscombe postpones.

Second, Anscombe does not say that the agent must be right in
thinking that the result may be brought about by the action, but
only that we can understand him thinking it. Here, Anscombe
illustrates the way in which what we can understand him thinking
is sensitive both to our context and knowledge, and to his:

Consider this case: ‘Why are you crossing the road?’ – Because there
will be an eclipse in July”. This answer, as things stand, needs fill-
ing in. And no kind of filling in that we shall accept without objection
would give that answer the role of a statement of intention. (I mean
something like ‘For six months before the eclipse that shop window is
having a lot of explanatory diagrams and models on display’).

(§22, p. 36)

Nevertheless, someone who lacked knowledge of the causation of
eclipses, and who believed, wrongly, that such events could be
influenced by human agency,

might well do something in order to produce an eclipse; and I suppose
the answer ‘Eclipse in July’ could perhaps have been understood as an

INTENTION WITH WHICH 123

expression of intention by the Dublin crowd who once assembled to
watch an eclipse, and dispersed when Dean Swift sent down his butler
with a message to say that by the Dean’s orders the eclipse was off.

(§22, p. 35)4

Anscombe’s observation is that the question of whether ‘There
will be an eclipse in July’ could be an expression of the inten-
tion with which something is done, depends for its answer on
whether we can make sense of someone thinking that an eclipse
is a later stage in proceedings of which her – or indeed someone
else’s – action P is an earlier stage. Given what we know about
the causes of eclipses, we know that there is no action P which
could fill this out. What the story of Dean Swift illustrates is that
this unintelligibility is sensitive to subject, time and context.
Dean Swift exploits the ignorance of the Dublin crowd – their
ignorance of what can bring about an eclipse – and his position
in the community as a person of ‘great power and authority’
(§22, p. 36). He correctly recognises that the hypothetical ‘If
Dean Swift orders the eclipse to be delayed then the eclipse will
be delayed’ is one that the crowd will accept. To paraphrase
Anscombe’s ‘vague and general formula’, they will see the delay
of the eclipse (the future state of affairs) as a possible later stage
in the proceedings of which Swift’s order ‘Delay the eclipse!’ is
an earlier stage. Emphasising this, Anscombe says: ‘on the one
hand, cases of scientific knowledge, and on the other hand cases
of magical rites, or of a vague idea of great power and authority
like Dean Swift’s, all come under this very vague and general
formula’ (§22, p. 36).

What Anscombe’s ‘vague and general formula’ introduces is the
idea of a calculative order. We ask Smith ‘Why are you A-ing?’
and he replies ‘In order to B’. For his answer to be acceptable – as
an answer to the question ‘Why?’ which gives the intention with
which he is A-ing – we must be able to make sense of him believ-
ing the hypothetical: if A then B.

This can all begin to sound highly psychological – as if it is
about minds not the mind. But, again, remember that intelligibil-
ity conditions are associated with the use of the concept: that is,
when we apply it, withhold it, recognise what others say and do

INTENTION WITH WHICH124

as falling under it: ‘The capacities picked out by “the mind” are
those for coherence and commitment in thought and language,
they are capacities characteristically exercised in going on intel-
ligently and intelligibly’ (Diamond, 1991, p. 5). The fact that
Anscombe is talking about an individual does not mean she is
describing his mind; she is rather describing an individual with a
view to showing what belongs to ‘the mind’.

DOING Y IN DOING X (§§23–28)

According to Anscombe, this ‘vague and general formula’ applies
in just the same way when descriptions are ‘wider descriptions of
what is happening’, as when they are descriptions of something
future. If you ask ‘Why are you standing on one leg?’ my answer
‘I am doing Yoga’ may well be perfectly intelligible; ‘I am writing
a book’ will need some ‘filling out’ to be so (e.g. ‘I find that doing
my Yoga poses helps me to stay focused while I write’).

This will help us to understand the order that Anscombe finds
in her famous example of the man operating a water pump. She
introduces this example with the question, ‘Let us ask: is there any
description which is the description of an intentional action, given
that one occurs?’ (§23, p. 37):

A man is pumping water into a cistern which supplies the drinking
water of a house. Someone has found a way of systematically con-
taminating the source with a deadly cumulative poison whose effects
are unnoticeable until they can no longer be cured. The house is
regularly inhabited by a small group of party chiefs, with their imme-
diate families, who are in control of a great state; they are engaged
in exterminating the Jews and perhaps plan a world war. – The man
who contaminated the source has calculated that if these people are
destroyed some good men will get into power who will govern well,
or even institute the Kingdom of Heaven on earth and secure a good
life for all the people; and he has revealed the calculation, together
with the fact about the poison, to the man who is pumping. The death
of the inhabitants of the house will, of course, have all sorts of other
effects; e.g., that a number of people unknown to these men will
receive legacies about which they know nothing.

INTENTION WITH WHICH 125

This man’s arm is going up and down, up and down. Certain
muscles, with Latin names which doctors know, are contracting and
relaxing. Certain substances are getting generated in some nerve
fibres – substances whose generation in the course of voluntary move-
ment interests physiologists. The moving arm is casting a shadow
on a rockery where at one place and from one position it produces a
curious effect as if a face were looking out of the rockery. Further, the
pump makes a series of clicking noises, which are in fact beating out
a noticeable rhythm.

(§23, p. 37)

Anscombe first notes that

our enquiries into the question ‘Why?’ enable us to narrow down our
consideration of descriptions of what he is doing to a range cover-
ing all and only intentional actions. ‘He is X-ing’ is a description of
an intentional action if (a) it is true and (b) there is such a thing as
an answer in the range I have defined to the question ‘Why are you
X-ing?’. But having applied this formula to rule out descriptions like
‘he is casting a shadow on the rockery’, there remain ‘a large number
of X’s for which we can readily suppose that the answer to the ques-
tion ‘Why are you X-ing?’ falls within the range.

(§23, p. 38)

This is where we found ourselves at the end of the last chapter:
with a ‘chaos’ of descriptions which ‘roughly outlined the area of
intentional actions’ (§18, p. 28). We had moved from a simple
distinction to a complex set of conditions, describing when an
answer gave a reason for acting.

Using her ‘vague and general formula’, Anscombe now arranges
the Xs that remain into series, where the series links descriptions
which fit into statements of the form: ‘I am doing Y in doing X’.
The man moves his arm; in moving his arm he is operating the
pump; in operating the pump he is replenishing the water supply;
in replenishing the water supply he is poisoning the people in the
house; in poisoning the people in the house …. Such a series could
be elicited using the question ‘Why?’, asked not once but repeti-
tively, reapplied over and over to the description in the answer given.

INTENTION WITH WHICH126

[A] I am moving my arm up and down.
Why are you moving your arm like that?

[B] I am operating this pump.
Why are you operating that pump?

[C] I am replenishing the water supply.
Why are you replenishing the water supply?

[D] I am poisoning the people in the house.
Why are you poisoning the people in the house?

[E] If we can get rid of them the other lot
will get in and things will be better.

Each answer [A]–[D] in the series is a description of his intentional
action and an expression of intention. Answer [E], Anscombe says,
marks a ‘break in the series of answers’ because now the answer
does not describe something he is doing in moving his arm but
only something that he is moving his arm in order to do.

this further description (e.g. to save the Jews, to put in the good men,
to get the Kingdom of Heaven on earth) is not such that we can now
say: he is saving the Jews, he is getting the Kingdom of Heaven, he is
putting in the good ones.

(§23, p. 40)

We can also imagine the series ending instead with a back-
ward-looking motive: ‘They humiliated me at the village fete’;
here it would be even clearer that the final answer was not such
that we could say ‘He is …’.

The question was: ‘is there any description which is the descrip-
tion of an intentional action, given that one occurs?’ (§23, p. 37).
First, Anscombe says, we can see that there is just one action: the
man moving his arm up and down. It is in moving his arm up
and down that he is doing all these things; he need do nothing
other than move his arm up and down in these circumstances to
be operating this pump, replenishing the water supply, poisoning
the people in the house: ‘The only distinct action of his that is in
question is this one, A’ (§26, p. 46).

INTENTION WITH WHICH 127

But, Anscombe says, it does not follow from this that ‘moving
his arm up and down’ is ‘the description’ of his intentional action.
Rather, she says, ‘there is one action with four descriptions, each
dependent on wider circumstances and each related to the next as
description of means to end’ (§26, p. 46). She adds that we can
speak of the four descriptions as being ‘swallowed up’ by the final
description which gives the ‘intention with which the act in its
other descriptions was done’ (§26, p. 46). So, we can say that he is
doing A, B, C and D; or we can just say, he is doing D. Anscombe
says we can ‘speak equally well of four corresponding intentions
or of one intention – the last term we have brought in the series’.

I want to connect this series of descriptions with the rela-
tion of ‘bruteness’ that we discussed at the end of Chapter 2.
It is important to be clear that when Anscombe speaks here of
‘dependen[ce] on wider circumstances’ she is not talking about
causal dependence. We can say: the relation of the description
‘moving his arm up and down’ to the description ‘poisoning
the inhabitants’ is that the former is brute relative to the lat-
ter (1958b, p. 172). This relation is one that exists only in the
context of human convention and normal proceedings – where
normal must again be specified relative to the ‘natural history of
mankind’. Furthermore, the descriptions in the series are deter-
mined by the agent; they are answers which fall within the range
of answers to the question ‘Why?’. If the pumping man is not
acting with the intention of poisoning the inhabitants, then that
description will not be added to the chain, even if his pumping is
as a matter of fact causing the inhabitants to be poisoned.

There are two further points to be noted about Anscombe’s reflec-
tions here. The first is a reminder of the sort of account of intention
Anscombe is offering and also a promissory note for Chapter 6; the
second concerns the ethical project described in Chapter 2.

First, Anscombe notes that:

nothing definite has to hold about how many terms we put between
A and D; for example, in the imagined case we did not put in a term
‘making the water flow along the pipes’, which yet would take its place
in the series if anyone thought of asking the question ‘Why?’ about it.

(§26, p. 47)

INTENTION WITH WHICH128

This shows quite clearly that Anscombe does not intend to
describe an order which involves ‘actual mental processes’ (§42,
p. 80). Anyone who is seeking to replenish a water supply by
operating an ordinary sort of pump must be doing so with the
intention that water will flow along the pipes, will arrive in the
water tank, etc. If he denied that he intended to make water flow
along the pipes by operating the pump, but still claimed to be
replenishing the water supply, we would need some ‘filling out’
which would have to involve his account of how, in his mind,
a pump worked. It seems better to say that the order belongs to
the concept ‘replenishing the water supply’ and, more gener-
ally, to the concept of intention. This is the promissory note for
Chapter 6: there we will investigate the idea that the order we
are describing here is somehow contained within the descrip-
tions we apply.

Concerning Anscombe’s ethical project, recall from Chapter 3
that Anscombe connects the Cartesian account of intention with
the idea that you can determine your intention by ‘making a little
speech to yourself: “What I mean to be doing is …”’ (Anscombe,
1961, p. 59). She associates this with ‘double think about double
effect’; it makes a mockery of that doctrine by offering ‘a marvel-
lous way … of making any action lawful’ (1961, p. 59).

Between §23 and §27, Anscombe makes a number of remarks
that are an attack on the Cartesian account of intention, and
which serve to provide a philosophy of psychology which can sup-
port an ethics of prohibition – one that is reliant on the doctrine
of double effect. The first thing to note is that Anscombe’s ‘vague
and general formula’ imposes intelligibility conditions on what
an agent can say in answer to the question ‘Why?’. In particular,
once an agent acknowledges application of the question ‘Why?’
to what he is doing under some description, by, e.g. giving an
answer to the question ‘Why are you operating that pump?’ which
falls within the specified range, this creates a demand for a further
reason. If he says ‘I am replenishing the water supply’ and he is
replenishing the water supply, then

the same question ‘Why?’ will have application to this action in its
turn. This is not an empty conclusion: it means that someone who,

INTENTION WITH WHICH 129

having so answered ‘To replenish the water supply’, is asked ‘Why
are you replenishing the water supply?’, he must not say e.g. ‘Oh,
I didn’t know I was doing that’, or refuse any but a causal sense of
the question.

In the case at hand it is also clear that ‘No reason’ or ‘I don’t know
why I did it’ will be quite unintelligible.

This fact immediately places fairly strict constraints on the sort
of lies that are possible, and on scope for disavowing or disowning
what one does. It is enough to undermine the Censor’s defence of
Truman:

[He argued] that Mr. Truman did not make the bombs himself, and
decide to drop them without consulting anybody; no, he was only
responsible for the decision. Hang it all, you can’t make a man respon-
sible just because ‘his signature is at the foot of the order’.

(Anscombe, 1957c, pp. 65–66)

If Truman says ‘To give the order to bomb Hiroshima’ in answer
to the question ‘Why are you signing that piece of paper’ he must
be prepared to answer the question ‘Why are you giving the order
to bomb Hiroshima?’. ‘No reason’ is clearly no answer.

These limits, however, still seem to leave space for ‘abuse’ of the
doctrine of double effect, which Anscombe spoke of in Chapter 2.
She imagines that the chain of answers given by the man operating
the pump take a different route:

[A] I am moving my arm up and down.
Why are you moving your arm like that?

[B] I am operating this pump.
Why are you operating that pump?

[C] I am replenishing the water supply.
Why are you replenishing the water supply?

[D] I’m just doing my job.
But the water is poisoned!

I know, I don’t care about that. I just want
my pay.

INTENTION WITH WHICH130

Anscombe notes that if he is being honest, then

although he knows concerning an intentional act of his … that it is also
an act of replenishing the water supply with poisoned water, it would be
incorrect, by our criteria, to say that his act of replenishing the house
supply with poisoned water was intentional.

(§25, p. 42)

Thus, if we are concerned with the question of whether this man
is a murderer, we must say that he is not. This observation, she
says, can make it natural to think that what makes the difference
between the case in which what happens is replenishing the water
supply and replenishing the water supply with poisoned water must
be ‘an interior movement’, since it is not anything that is observ-
ably going on in the situation. And this is where the idea that he
might direct his intention in this way might arise.

[N]ow if intention is an interior movement, it would appear that
we can choose to have a certain intention and not another, just by
e.g. saying within ourselves: ‘What I mean to be doing is earning
my living, and not poisoning the household’ … The idea that one
can determine one’s intentions by making such a little speech to
oneself is obvious bosh. Nevertheless the genuine case of ‘I didn’t
care a tuppence one way or the other for the fact that someone had
poisoned the water’ … does appear to make it very difficult to find
anything except a man’s thoughts … to distinguish the intentional
poisoning from poisoning knowingly when this was nevertheless not
the man’s intention.

(§25, p. 42)

Anscombe addresses two parts of this line of thought. First, she
answers that ‘there can be a certain amount of control for truth-
fulness of the answer’ (§25, p. 43). This control comes not from
an ability to ‘see his mind’, but from the fact that what he says
in answer to the question ‘Why?’ must be intelligible, and what
he says can be made unintelligible if it fails to fit with what he is
observably doing. In the case we are considering:

INTENTION WITH WHICH 131

It is … necessary that it should be his usual job …; and he must not do
anything, out of the usual course of his job, that assists the poisoning
and of which he cannot give an acceptable account. E.g. suppose he
distracts the attention of one of the inhabitants from something about
the water source that might suggest the truth; the question ‘Why did
you call him from over there?’ must have a credible answer other than
‘to prevent him from seeing’; and a multiplication of such points
needing explanation would cast doubt on his claim not to have done
anything with a view to facilitating the poisoning.

(§25, p. 43)

Anscombe points out that we have ‘stock way[s] of bringing out
pretences’, often ‘met with in literature’: we can try to elicit reac-
tions by saying things like ‘Well you won’t care that the poison is
old and won’t work’ or ‘Then you won’t be claiming a share in a
great sum with which someone wants to reward the conspirators’
(§27, p. 48). We investigate how committed a worker he really is,
whether he knows the plotters, what his political beliefs are, and
so forth. We can employ the methods of ‘psychological detectives’
(§27, p. 48). All this contributes to a great store of evidence that
he is lying if he is.

Anscombe thinks that there may nevertheless be a point at
which ‘a man can say “This is my intention”, and no one else can
contribute anything to settle the matter’ (§27, p. 48). This, how-
ever, does not imply that a man who says ‘This is my intention’ can
settle that it is his intention by ‘making a little speech to himself ’.

Suppose that the man imagined, who says ‘I was only doing my usual
job’, might find this formula and administer it to himself in the pres-
ent tense at some stage of his activities. However, if he does this, we
notice that the question arises: with what intention does he do this?
This question would always arise about anything which was deliber-
ately performed as an ‘act of intending’. The answer in this case might
be ‘So that I didn’t have to consider whose side I am on’. Thus the
interior performance has not secured what you might have thought,
namely that the man’s action in pumping the water is just doing his
usual job; it is itself a new action.

(§27, p. 47)

INTENTION WITH WHICH132

Suppose we ask ‘Why are you operating that pump?’ An answer: ‘I
was going to refuse but then I had the thought: “I am only doing
my usual job!”, and so I started pumping’. This clearly gives a
mental cause and not a reason.

NON OBSERVATIONAL KNOWLEDGE? (§§29–33)

Using Anscombe’s ‘vague and general formula’ we have been able
to describe an order that is there in the descriptions which fall
within the class outlined in Chapter 4 – the class of descriptions
of intentional actions. We have also seen how the formal require-
ments imposed by that order can begin to undermine the idea that
‘if we want to know a man’s intentions it is into the contents of
his mind and only into these that we should enquire’ (§4, p. 5).

We now come to §§33–45, which contain a discussion of
Aristotle’s account of practical reasoning. Before we turn to that
discussion I want to ask a more fundamental question about those
sections: what are they for? To see why this question arises, look
ahead at what Anscombe says at the end of her eleven-paragraph
discussion of practical reasoning:

Aristotle’s account [of practical reason] describes an order which
is … the same order as I arrived at in discussing what ‘the intentional
action’ was, when the man was pumping water. I did not realise the
identity until I had reached my results; for the starting points for my
enquiry were different from Aristotle’s, as is natural for someone writ-
ing in a different time.

(§42, p. 80)

This makes it sounds as if there is nothing in that discussion that
takes us beyond what is already contained in §§23–27.

In one sense that is quite right. The sections on practical rea-
soning redescribe the order that is already specified in the ‘vague
and general formula’ of §22. In §47, Anscombe says:

[T]he term ‘intentional’ has reference to a form of description of events.
What is essential to that form is displayed by the results of our enquiry
into the question ‘Why?’

(§47, p. 84)

INTENTION WITH WHICH 133

And this leads to an interpretative puzzle. If it is right that
Anscombe has already described what is essential to that ‘form
of description’ by the end of §27, she was already then in a posi-
tion to make this statement about the character of the concept of
intention. What, then, is the discussion of practical reasoning for?

It may be tempting in light of this to pass over these para-
graphs. Because we know the insight of §42 – the identity of the
order displayed by the question ‘Why?’ and the order described
in practical reasoning – we need not be too concerned with them.
Indeed, these passages have received very little discussion in the
literature. If we are in the mood to see Intention as a ‘first draft’,
we might assume that Anscombe left her working-out here, and
that future revisions might have relegated this discussion to the
footnotes. We perhaps see Anscombe attempting something like
an ‘appeal to authority’; she thinks what she says might gain more
traction if she can claim its genesis in Aristotle. Or, more gener-
ously, we might take a scholarly interest in her interpretation and
appropriation of Aristotle.

It would be a mistake to take this attitude. To see why we need
to understand the infamous challenge posed by the interlocutor,
in the intuitive voice, at §28. It is in response to that challenge
that Anscombe introduces the topic of practical reasoning. As
we will see, the challenge cannot be met using the discussion of
the man pumping, precisely due to its ‘different starting point’.
The description of the man pumping makes sense only if the very
thing that the interlocutor finds unintelligible is already taken for
granted. This is because it relied, in setting out the order that is
there in intentional action, on the idea that when the man says,
e.g., ‘I am replenishing the water supply’ this is both an expres-
sion of his intention and a description of his intentional action.
And this is precisely what the interlocutor – when he interrupts
at §28 – finds unintelligible.

At the end of her discussion of the man operating the pump,
Anscombe says: ‘We must now look more closely into the for-
mula which has so constantly occurred in this investigation:
“known without observation”’ (§28, p. 49). When she says it has
‘constantly occurred’ she means: it has been taken for granted

INTENTION WITH WHICH134

in the previous discussion. In §8 Anscombe introduced the idea
that ‘the class of intentional actions is a sub-class of [the class
of things known without observation]’ (p. 14), but said nothing
there to explain or elucidate that point. It is, I said, introduced
in the descriptive voice. As we noted, the ‘clarification’ of the dis-
tinction between intentional and unintentional actions – which
occurs between §5 and §19 – began with the class of things a man
knows without observation, and then worked to make distinc-
tions within that class.

‘We must now look more closely …’, to recall the discussion in
Chapter 2, signifies the need for a philosophical account, a need
which arises in response to an objection from the interlocutor that
the description of what we say and do cannot be correct. The
objection is the following:

‘Known without observation’ may very well be a justifiable formula for
knowledge of the position and movement of one’s limbs, but you have
spoken of all intentional action as falling under this concept. Now it
may be e.g. that one paints a wall yellow, meaning to do so. But is it
reasonable to say that one ‘knows without observation’ that one is
painting a wall yellow? And similarly for all sorts of actions: any action
that is, that are described under any aspect beyond that of bodily
movement.

(§28, p. 50)

Anscombe immediately answers:

My reply is that the topic of an intention may be a matter on which
there is knowledge or opinion based on observation, inference, hear-
say, superstition or anything else that knowledge or opinion are ever
based on; or again matter on which an opinion is held without any
foundation at all. When knowledge or opinion are present concerning
what can be the case, and what can happen – say Z – if one does cer-
tain things, say ABC, then it is possible to have the intention of doing
Z in doing ABC; and if the case is one of knowledge or if the opinion is
correct, then doing or causing Z is an intentional action, and it is not
by observation that one knows one is doing Z.

(§28, p. 50)

INTENTION WITH WHICH 135

In giving this answer Anscombe describes what it is for a person
to have knowledge of what he is doing. The reply uses the ‘vague
and general formula’ we have already introduced, and adds a cor-
rectness condition: if he is correct ‘concerning what can be the case
and what can happen’, then in doing ABC with the intention of
doing Z he knows, not by observation, he is doing Z.

This answer is not provisional; it would be better to say, as she
says earlier of her account of expressions of intention: it is ‘not
false but rather mystifying’ (§3, p. 7). It is ‘mystifying’ in the sense
that we have deep objections to accepting what Anscombe is say-
ing. Anscombe brings out the objection by giving the interlocutor
a chance to pose a challenge to her reply, this time connecting the
difficulty explicitly with the fact that what is said to be known
‘in intention’ (§32, p. 57) is also something that can be known
‘by observation’:

The difficulty however is this: What can opening the window be except
making such-and-such movements with such-and-such result? And in
that case what can knowing one is opening the window be except know-
ing that that is taking place. Now if there are two ways of knowing here,
one of which I call knowledge of one’s intentional action, and the other
of which I call knowledge by observation of what takes place, then
must there not be two objects of knowledge? How can one speak of
two different knowledges of exactly the same thing? It is not that there
are two descriptions of the same things, both of which are known, as
when one knows that something is red and that it is coloured; no, here
the description, opening the window, is identical, whether it is known
by observation or by its being one’s intentional action.

(§29, p. 51)

It is exceptionally difficult to get a grip on this objection and it is
not clear that anyone who has written on this part of Anscombe’s
book has found a satisfactory way of parsing the interlocutor’s con-
cern.5 An easy way out for an interpreter is to say: well, of course
the objection is inchoate; it is precisely Anscombe’s point that this
‘natural’ objection is misguided. So it is bound to resist coherent
summary. There is something right about this. However, if we are
to make sense of what the discussion of practical reason is for we do

INTENTION WITH WHICH136

need to get to grips with what it is that – perhaps inchoately – the
interlocutor finds objectionable in Anscombe’s first answer.

The key idea concerns ‘objects of knowledge’. Anscombe
repeats it later: ‘if there are two knowledges – one by observation,
and the other in intention – then it looks as if there must be
two objects of knowledge’ (§32, p. 57). It looks like a solution
to the puzzle will make clear how there can be ‘two knowledges’
and ‘one object of knowledge’ (compare §48, p. 89). Indeed, I
want to propose that this is what Anscombe’s account of practical
reasoning is meant to deliver.

What is difficult to resolve in these passages, and in the various
forms of the interlocutor’s worry, is how – or indeed whether – the
impression that ‘two knowledges’ implies ‘two objects of knowledge’
stems from a general point about how to ‘count’, as it were, knowl-
edges or whether it is connected to particular features of ‘knowledge
in intention’, namely, that it is ‘without observation’. I suspect that
ultimately the two worries come from the same false assumption
about knowledge, an assumption that Anscombe aims to dislodge,
however, I here admit defeat in trying to bring the connection into
clear focus.

The interlocutor sometimes seems to be making a quite general
point about knowledge. The point being: you can only speak of
‘two knowledges’ where you have two objects of knowledge. To
see why you might want to say this, recall the case we looked at
earlier, where Jones was sawing a plank. Anscombe said then, in
her descriptive voice: ‘it is important to notice that a man may
know that he is doing a thing under one description and not
under another … So to say that a man knows he is doing X is to
give a description of what he is doing under which he knows it’
(§7, p. 12). Now, if we took ‘object’ in ‘object of knowledge’ to
mean thing, then the interlocutor would have to be insisting that
Jones could only have ‘one knowledge’ of what he is doing since
all that is happening in the situation is, e.g. he is moving his arm
back and forth. This would be a very odd idea. What would be less
odd would be to say that he would at least have to be able to know
this one thing – what is happening – under two descriptions. We
have already seen this conception of ‘object’ in Chapter 3; ‘object’
in ‘object of knowledge’ should be understood intentionally,

INTENTION WITH WHICH 137

like ‘object’ in ‘object of sight’. So, we can speak of two objects of
knowledge if Jones knows he is sawing a plank and making a pile
of sawdust, even though these are two descriptions of one thing
he is doing. In the passage quoted above, Anscombe, says ‘it is not
as if there are two descriptions of the same thing, both of which
are known … ; no, here the description, opening the window, is
identical’ (§29, p. 51).

When we gloss the objection that way, nothing seems to turn
on Anscombe’s saying that knowledge in intention is without
observation. The objection is just: if you’re going to talk about
‘two knowledges’, you better have two objects of knowledge, but
if what I know ‘in intention’ and what I know ‘by observation’ is a
single thing under a single description why talk of ‘two knowledges’
unless what you really have is two subjects (cf. §32, p. 57)?

The second version of the interlocutor’s worry does turn
on the fact that the knowledge ‘in intention’ is supposed to be
without observation. Anscombe discusses this in §29. She says
that she has ‘come up with the formula “I do what happens”’
as a way of trying to respond to the interlocutor’s worry. This
would solve the ‘two object’ puzzle because the object of knowl-
edge ‘in intention’ – a description of what I do – and the object
of observational knowledge – a description of what happens
would match (one object of knowledge), but we would still have
‘two knowledges’. We would have ‘two knowledges’ because, as
Anscombe’s first reply (§28, p. 50) makes clear, knowing what I am
doing has a different form from knowing what is happening. But
Anscombe says that people find this ‘paradoxical’ because: ‘what
happens must be given by observation; but I have argued that my
knowledge of what I do is not by observation’ (§29, p. 53). So,
on this second version of the interlocutor’s worry, the problem is
not that there is only one available object of knowledge, but that a
description of what I do, even if it happens to be a description of
what is happening, can’t count as knowledge.

Both versions of the objection point toward the idea that
knowing what I am doing can’t be knowing what is happening.
On the first version, it can’t be because what is happening is
already known under that very description by observation; on
the second it can’t be because what is happening couldn’t be

INTENTION WITH WHICH138

known except by observation. A solution to the first would show
how the very same description could be known in a way that
made it right, still, to speak of two knowledges. A solution to
the second would show how a description of what one is doing,
when it was also a description of what was happening, would
count as knowledge. I suggest that Anscombe aims to discharge
both these requirements through the account that she gives of
practical reasoning.

Anscombe identifies two, connected, features of our ‘natural’
thinking about practical reasoning as lying behind the interloc-
utor’s objection: first, the view that it is ‘reasoning toward the
truth of a proposition which is supposedly shown to be true by
the premises’ (§33, p. 58); second, the view that the action that
arises is compelled or necessitated by the reasoning. Once these
features have been rejected, Anscombe can secure an account of
a kind of reasoning which is not reasoning toward the truth of a
proposition, and she can describe a kind of knowledge which does
is not connected with the belief or judgment that such-and-such is
the case. She calls that knowledge ‘practical knowledge’.

I will present the account of practical reasoning now, and in
Chapter 6 look at what it delivers: practical knowledge.

PRACTICAL REASONING AS ‘ORDINARY
REASONING’ (§§33–34)

When a philosopher introduces the term ‘practical reasoning’
she often means to mark an intuitive distinction with reasoning
which is ‘theoretical’. The intuitive distinction can be given by
an example: when a man reasons that he should spend more time
with his children because being a good father demands it of him,
his reasoning is practical. When he reasons that the weather will
be unsettled because there is an area of low pressure approach-
ing, his reasoning is theoretical. The difficulty for a philosopher
is to give some account of difference that this ‘intuitive distinc-
tion’ marks.

Anscombe says that the character of practical reasoning has
been ‘obscured’ by the fact that ‘it is commonly supposed to be
ordinary reasoning leading to such a conclusion as: “I ought to

INTENTION WITH WHICH 139

do such and such”’. By ‘ordinary reasoning’ she means ‘reasoning
toward the truth of a proposition which is supposedly shown to be
true by the premises’ (§33, p. 58).

Anscombe highlights some of the difficulties that arise when
one makes this supposition. To begin, it implies that practical
reasoning is to be distinguished by its practical or ethical subject
matter alone. The conclusion of a piece of practical reasoning is
a proposition with an ethical or practical content, e.g. ‘I should
do X’ or ‘It would be good for me to do X’. Anscombe objects
that a difference in subject matter is not enough to ‘justify us in
speaking of a special sort of reasoning’ (§33, p. 58). As she puts
it, ‘one might easily wonder why no one has ever pointed out
the mince pie syllogism’ (§33, p. 58), the conclusion of which
would, presumably, be a proposition about mince pies. Her
point here is that if subject matter alone is enough to mark out
kinds of reasoning, we should not stop at practical reasoning,
but may as well, for each possible topic one might reason about –
for example, mince pies – speak of a distinct kind of reasoning.
The fact that ‘ethics is of importance to human beings in a way
that mince pieces are not … cannot justify us in speaking of a
special sort of reasoning’ (§33, p. 58). Love, politics, religion,
children, sex, food, and health are of importance to human
beings, but we wouldn’t speak of a ‘special sort of reasoning’
applicable to each.6

Anscombe notes a second, more serious, problem with the idea
that the conclusion of a piece of practical reasoning is a proposition
about what one ought to do. Anscombe points out that when one
draws a conclusion about what suits me, or what it would be good
to do, ‘nothing seems to follow about doing anything’ (§33, p. 59).
Anscombe illustrates the sense in which ‘the reasoning does not
compel any action’:

It is obvious that I can decide, on general grounds about colouring
and so on, that a certain dress in a shop window would suit me very
well, without its following that I can be accused of some kind of incon-
sistency with what I have decided if I do not thereupon go in and buy
it; even if there are no impediments, such as shortage of cash, at all.

(§33, p. 59)

INTENTION WITH WHICH140

It is quite rational, reasonable, intelligible that I should draw the
conclusion that it would be good or suitable to do something and
yet not do it. Even if there are no impediments whatsoever, when
I draw it I need not act.

We can, I think, connect this view of practical reasoning with
the deontological account of ethics that we saw Anscombe set her-
self against in Chapter 2. Recall that she insisted that the idea of
a ‘moral ought’ was quite unintelligible outside a Judeo-Christian
belief-system. The need for a special moral ought must come in
part from the idea that we need an ‘ought’ in the conclusion of a
piece of practical reasoning such that, should the agent draw the
conclusion, he must act. The idea of a moral ought is the idea of
such an irresistible norm.

One way to try to get a piece of practical reasoning to ‘compel’
action, while retaining the formal character of theoretical reason-
ing is to add a universal premise in imperative form which then
yields a conclusion (C) in that form.

1 Do whatever is good for a human adult!
2 Getting thirty minutes exercise a day is good for a human

adult.
C Get thirty minutes exercise a day!

Anscombe remarks that ‘[t]he syllogism in the imperative form
avoid [the] disadvantage [that it does not compel action] … but
suffers from the disadvantage that the … universal premise is an
insane one, which no one could accept for a moment if he thought
what it meant’ (§33, p. 59).

Consider:

1 Do whatever is conducive to not having a car crash!
2 Such-and-such is conducive to not having a car crash.
C Do such-and-such!

Anscombe points out:

[T]here are usually a hundred different and incompatible things con-
ducive to not having a car crash; such as, perhaps, driving into the

INTENTION WITH WHICH 141

private gateway immediately on your left and abandoning your car
there, and driving into the private gateway immediately on your right
and abandoning the car there.

(§33, p. 59)

So too, ‘what is good for a human adult’ is: foster loving friend-
ships and relationships, spend time with one’s family, relax,
work, enjoy oneself, contemplate the meaning of life …. The
problem is that any universal premise that is not negative will
be ‘insane’. No matter how specific or constrained or hedged or
particular we make ‘X’, any imperative ‘Always do X!’, if applied
in particular cases, will impel actions that are plainly irrational.
Anscombe remarks that ‘only negative general premises can hope
to avoid insanity of this sort [but] these, even if accepted as prac-
tical premises, don’t lead to any particular actions … but only
to not doing certain things’ (§33, p. 61). If we restrict practical
reasoning to reasoning which leads to inaction, then we will need
to give some other account – with another name – of reasoning
which leads to action.

Anscombe rejects both the idea that the conclusion of practical
reasoning is a proposition and that the premises of a piece of practi-
cal reasoning necessitate the conclusion. To paraphrase Anscombe’s
remark about expressions of intention and prediction: the distinc-
tion is intuitively obvious, but cannot be left so (§3, p. 6).

THE IDEA OF LOGICAL COMPULSION (§33)

Anscombe argues, in her famous paper ‘Practical Inference’
(1974a), that part of the difficulty we have in defining practical
reasoning lies in a failure to understand the character of theoretical
reasoning. She thinks that we must insist that ‘there is a difference
of form between reasoning leading to action and reasoning for the
truth of a conclusion’ (§33, p. 60), but that our failure properly to
characterise the form of the latter kind of reasoning leaves us una-
ble to see where this difference might lie. If theoretical reasoning
is defined as a formal relation between premises and conclusion –
the truth of the premises imply the truth of the conclusion – then
it seems that saying that practical reason has a different form will

INTENTION WITH WHICH142

imply that it is not really reasoning at all. Hence, on pain of los-
ing the idea that practical reasoning is genuinely reasoning, we are
compelled to differentiate it by subject matter and by trying to
build in a compulsion to act on the conclusion, in addition to a
compulsion to believe it. As we will see, this error is connected
with the tendency to treat the topic of intention as psychological
rather than logical: as about minds not the mind.

Anscombe makes space for a difference in form which is con-
sistent with thinking of practical reasoning as reasoning proper, by
first suggesting that the difficulties about how the premises of a
piece of practical reasoning can compel action can be relocated to
the theoretical realm. The idea of a logical compulsion is as much
a problem if we are trying to understand the relationship between
p, if p then q, and the belief that q, as it is if we are trying to under-
stand the relationship between the conditional and my decision to
bring it about that p.

The transmission of intention and transmission or belief should be
put side by side; … there is no such thing as the transmission’s being
‘logical’ in the sense that the ‘necessity’ of the truth-connexion has
an analogue in a ‘logical compulsion’ to be in one psychological state
rather than another.

(1974a, p. 130)

The puzzle in the practical realm can be raised in the theoretical:
how, we can ask, can p, and if p then q, compel the belief that q?

Anscombe’s suggestion is that we separate out the hypotheticals –
if p then q – and the use of the hypotheticals by a human being
who has a particular aim which relates to q. A man might have
a number of different aims in relation to q. His aim may be to
find out whether q (or ~q), or it may be to make it the case that
q (or ~q). If the hypothetical is used with aims that relate to find-
ing out whether q is true, the reasoning will be theoretical; if used
with aims that relate to bringing q (or ~q) about, the reasoning
is practical. Anscombe points out that there are other uses too:

These hypotheticals might of course be used to make a threat or offer
a warning … There is also a use in seeking an explanation; we have it

INTENTION WITH WHICH 143

given that r, and the hypotheticals suggest p, which we will suppose is
something we can check for truth. If p turns out true, it may perhaps
explain r.

(1974a, p. 129)

I might also aim to demonstrate the conditions under which
q, or to deceive others into thinking that q, or to set out the case
for doing q, and so on and so on. Anscombe’s observation is that
the hypotheticals, which record our collective and individual
knowledge, opinion, belief, prejudices, about what follows from
what, can be used by an individual who has some end in view,
be it practical, theoretical, or explanatory. The subject matter of
the hypotheticals might concern anything whatsoever, even claims
about mince pies.

If Anscombe is right, the relation between the reasoning and
the conclusion is not to be located in the individual psychology
of a human mind, nor is what is distinctive about practical rea-
soning located in its topic. Though, she acknowledges, ‘[there]
would be no point in proof patterns, if they were never to be
plugged into believing minds, if nothing were ever asserted’ and
‘equally no point in patterns of practical inference if nothing
were ever aimed at’, it is an error to put ‘the wanting or intend-
ing or believing into the description of the inferences’ (1974a,
p. 139). Just as we can display the formal character of theoreti-
cal reasoning without reference to belief, so too we can display
the formal character of theoretical reasoning without reference
to desire.

By relocating the psychological states of an individual – their
beliefs, desires, wants – outside the logical form of practical and
theoretical reasoning, which is ‘the truth-connexions of p, if
p then q, and q; and of not-p, only if p then q, and not-q’ (1974a,
p. 139), Anscombe is able to present a formal difference between
practical and theoretical reasoning which does not contradict the
claim that practical reasoning is reasoning proper. When I reason
practically I reason from a desire – to get o or to make it the case
that p – to action that I believe will get me o or bring p about,
via hypothetical propositions which concern what follows from
what. The object of my desire will figure in the first premise in

INTENTION WITH WHICH144

a piece of practical reasoning but not my desire for that object.
Rather, my having that desire for that object is what makes this
calculation – the one that sets out how to achieve that object –
have a ‘point’ for me. If I have no desire to become an accountant,
then patterns of inference which begin, ‘Doing XYZ is a way to
become an accountant’ will probably hold no interest for me. If I
have a specific desire not to become an accountant – if, for exam-
ple, five generations of my family have had such a career and I am
desperate not to be railroaded into joining the family firm – then
such patterns may lead me to avoid doing certain things, for exam-
ple, signing up for particular modules during my degree.

Because there are usually many ways in which one can attain
what one wants, it is rare for a particular action to be compelled
by one’s desire. For example, if I want to poison the inhabit-
ants, I can replenish the water supply with poisoned water, but
I could also get a job in the bakery and supply them with arse-
nic-laced macaroons. And even if replenishing the water supply
with poisoned water was the only way to get them poisoned, I
could contaminate the source at the well or by meddling with the
supply at the pump.

There is only one case in which we might wish to speak, and
might speak intelligibly about practical reasoning compelling
or ‘necessitating’ a particular action (1974a, p. 120) and having
such a case in view should, I think, help us to see what Anscombe
insists: that the idea of necessitating actions has no place in an
account of practical reason. Such cases are those in which the
person reasoning recognises one and only one way of achieving
what he wants. Here, the first line in the reasoning will not be
‘If p then q’ but ‘q only if p’.7 Anscombe observes that these cases
are extremely rare (§44, p. 81)8 and Candace Vogler offers an
example to show how absurd it is to suppose that this kind of
case is ordinary:

[W]hy would we suppose that we ever had discovered the ‘indispen-
sably necessary means’ to attaining an end? Well if the means that ‘lie
within [a man’s] power’ are sufficiently constrained – say, if he and his
buddy are handcuffed to a bomb that will go off in two minutes, and
one of them must pick the handcuff lock if they are to free themselves,

INTENTION WITH WHICH 145

and the only object available for the lock-picking is a bit of wire pro-
truding from one man’s shoulder … that he must yank free with his
teeth and drop into his buddy’s palm if the buddy (who has the requi-
site locksmith skill) is to pick the lock, and the man stuck with the wire
is determined that the two of them will free themselves, then his only
choice is to yank the wire out, and so on – then he confronts indispen-
sably necessary means to his end.

(Vogler, 2002, p. 154)

WANTING (§§34–40)

This is the background for Anscombe’s insistence that ‘the role of
“wanting” in practical syllogism is quite different from that of a
premise’ (§35, p. 66). Rather than being a premise, she argues that
‘whatever is described in the proposition that is the starting-point
of the argument must be wanted in order for the reasoning to lead
to any action’ (§35, p. 66).

To make the parallel case: the question ‘Should I believe p?’ does
not appear in a piece of theoretical reasoning, but is the starting-
point of an argument if it is to lead to any belief. Anscombe says
that ‘the conclusion [of a practical syllogism] is an action whose
point is shewn by the premises, which are now, so to speak, on
active service’ (§33, p. 60); and, we can add, the conclusion of a
theoretical argument is a belief whose justification is shown by the
premises, which are now on active service.9

This can seem like a technical point, but we can now see that
it is deeply connected with the need to exclude the psychological
from the formal character of practical reasoning – and hence from
the account of intention – and with Anscombe’s earlier concern
to distinguish two sorts of wanting in relation to action: the sort
that is a mental cause and the sort that gives the object of one’s
desire (§§10–11). If we begin with a psychological premise, then
we build into the reasoning the idea of psychological compulsion,
of reasoning as ‘transmission’ of intention or belief (Anscombe,
1974a, p. 130). We then end up with the idea that the hypo-
theticals exert a pseudo-mechanical logical compulsion, whether
toward belief or action, or that the inferences hold between states
of mind, carrying justification along a chain of mental states.10

INTENTION WITH WHICH146

This will be to reinstate precisely the kind of psychological-causal
account of intention that Anscombe is seeking to overturn, and
which she calls ‘Will-o’-the-wisp’.11 Freed from this psychological
context the idea of an ‘order’ ceases to be ‘a historical order of
actual consideration’ but rather a formal order: ‘a man’s considera-
tions leading to an action can be arranged in an order that displays
a progress from something mentioning an end to the particular
action adopted’ (Anscombe, 1974a, p. 142). When no such order
can be discovered using the question ‘Why?’ or by careful observa-
tion, we will not recognise what is happening as falling under the
concept intention. As Anscombe says: ‘if Aristotle’s account were
supposed to describe actual mental processes, it would in general
be absurd’ (§42, p. 80).

In this context, Anscombe launches an attack on the ‘mon-
olithic conception of desire, or wanting, or will’ which she says
is ‘a great fault’ in ‘modern philosophy of the Anglo-American
tradition’ (Anscombe, 1993, p. 154). The kind of ‘wanting’
that Anscombe is isolating here is, she remarks, nothing like the
empiricist idea according to which ‘wanting would be an internal
impression’ (§40, p. 77).

Here we return to the distinction we made at §11. The question
‘What do you desire?’ could be answered by describing a feeling.
However, there is version of the question that is not asking for
a ‘mental cause’ or an inner impression – that cannot be parsed
‘What do you feel?’ – but which asks for an object of desire in the
context of an action.

The question ‘What do you want?’ was not a question out of the blue,
like ‘What are the things you want in life?’ asked in a general way at
the fireside. In context, it is the question ‘With a view to what are you
doing X, Y, and Z?’, which are what he is doing. That is to say, it is a
form of our question ‘Why?’ but with a slightly altered appearance.

(§35, p. 63)

This kind of wanting, then, shares the order that belongs to
practical reasoning. We can ask ‘Why do you want X?’ and the
intelligibility conditions set by the ‘vague and general formula’ will
apply here too.

INTENTION WITH WHICH 147

Anscombe points out that the kind of wanting which is the
starting point for practical reasoning has some formal con-
straints on what it can take as its object – constraints not shared
by the sort of wanting that might be given as a mental cause.
These constraints could be thought of as limits on the ‘subject
matter’ of practical reasoning without it now looking as if we are
‘distinguishing [the] reasoning by its subject matter’ (§33, p. 59).
Anscombe begins with an ‘absolute restriction’ (§37, p. 70): ‘prac-
tical reasoning is essentially concerned with “what is capable of
turning out variously”’ (§33, p. 60). I cannot want – in the sense
that would make the object of my wanting the starting point for
practical reasoning – something that is not contingent, in that
sense. (Contrast the mental cause: ‘I am crying because I want to
have been prettier’.)

After that, she says, there are ‘some relative’ restrictions (§37,
p. 70). The restrictions relate to the relation between what is
wanted and what is desirable, and again must be specified in rela-
tion to the ‘natural history of mankind’. In order to make sense
of a person wanting something we need to be able to see what
it is about it that he finds desirable. There are obviously enor-
mous differences between what I will find desirable, what you will
find desirable, what a Nazi, Catholic, cat, architect, hedge fund
manager, farmer and aid worker will find desirable. However,
Anscombe says, anything that is wanted should come under one
of Aristotle’s three terms: ‘should’, ‘suits’, ‘pleasant’ (§35, p. 63).
If I say ‘I want X’, then for what I say to be intelligible, I must be
able to say ‘what for?’ where this terminates in an account of why
‘X’ is suitable, pleasant, or fitting.

NOTES

1 This kind of question will be familiar to readers of Wittgenstein’s later
work. Compare also Anscombe’s remarks on wishing:

One can imagine the existence of a people whose language did
not include the expression of a wish that things had been other-
wise. It would be possible to formulate the wish in their language
by using their expressions for wishing and their past tense;

INTENTION WITH WHICH148

yet it might be that to them this sounded incomprehensible, or like
mere bad grammar. If such people existed they would seem to be
psychologically different from us.

(1950, p. 116)

This example could be taken to show that the past-tense use of wish
is not essential to the concept of a wish: we are not inclined to say of
these people that they do not wish. The sense in which these people
are ‘psychologically different’ is not that there is a mental state we
can get into which they can’t, but that they won’t be wistful about
the past or have regrets about it in some of the ways that we do: ‘I
wish it had been otherwise’ is an important expression for us. See
Cora Diamond (1966), for discussion of the connection between this
passage and Anscombe’s claim that talk of moral obligation without
God is absurd.

2 Of course, he could answer ‘I am building a squirrel house’, but here
the appearance of speaking in the present-tense really does belong to
superficial grammar. The temporal marker ‘during the summer break’
is carried over from the question. Jones is clearly not building a squir-
rel house – if he is then, unless he is a serial squirrel-house builder,
his statement that he will build one over the summer break looks
quite obscure.

3 In this context Anscombe briefly considers whether the distinction
between voluntary and involuntary actions could be drawn some other
way. For example, we might say that voluntary actions can be foreseen
or that voluntary actions can be commanded (§20, pp. 32–33). As with
all ways of drawing distinction, these work only if involuntary actions
cannot be foreseen or commanded, and if all voluntary actions can be
foreseen or commanded. Anscombe suggests neither of these claims
is true. I will not pause over this argument here as we will look at predic-
tion and command in Chapter 7.

4 The following extract is from The Works of the Rev. Jonathan Swift:

One day Swift observed a great rabble assembled in a large space
before the deanery door in Kevin street, and upon inquiring the
cause of this, was told it was to see the eclipse. He immediately
sent for the beadle, and gave him his lesson what he should do.
Away ran Davy for his bell, and after ringing it some time among

INTENTION WITH WHICH 149

the crowd, bawled out, O yea, O yea, all manner of persons con-
cerned, are desired to take notice, that it is the dean of St Patrick’s
will and pleasure, that the eclipse be put off till this hour to mor-
row. So God save the king, and his reverence the dean. The mob
upon this notice immediately dispersed; only some, more cunning
than the rest, swore they would not lose another afternoon, for
that the dean, who was a very comical man, might take it into his
head to put off the eclipse again, and so make fools of them a
second time.

(Nichols, 1808, p. 416)

5 For some contrasting accounts and discussion of the difficulties see,
e.g. Haddock (2011), Moran (2004), Hursthouse (2000).

6 See Vogler for further discussion:

[W]e often treat the term ‘practical’ as marking out a region of
topics about which one can think, and practical reason as what is
at work when one thinks about such topics. The region of practical
topics consists in topics pertaining to action in general, with con-
siderations pertinent to one’s own doings as a kind of subdivision.
The latter are of special interest since they are most likely to be
brought to bear on what one does.

(Candace Vogler, 2002, p. 27)

7 This is the form that Von Wright (1972) considers (‘Unless I do A I shall
not attain E’).

8 Case (b) is of this kind: ‘A man has an end in view, e.g. to eat only
wholesome food, is always confronted with only one wholesome dish,
and recognizing it as a kind of food that is wholesome, he takes it and
not others’ (§44, p. 81.)

9 Anscombe points to Aristotle’s distinction between ‘classroom’ or
‘idle’ syllogisms, and syllogisms proper (§33, p. 60).

10 Of course, philosophers who don’t like the idea of a logical compulsion
will want to make the compulsion causal. For Anscombe, this view is
already ruled out given her argument about mental cause (see Chapter 4).
See also ‘Practical Inference’ (1974a, pp. 111–112) for her discussion of
Davidson’s attempt to give such an account. Anscombe says his solu-
tion ‘lacks acumen’ (p. 110).

11 Anscombe, ‘Practical Inference’ (1974a, p. 140).

INTENTION WITH WHICH150

SUGGESTED READING

Anscombe, G. E. M., (1974a). ‘Practical Inference’. From P. A. Schilpp and L. E. Hahn
(eds), The Philosophy of Georg Henrik von Wright. Library of Living Philosophers
Series Volume XIX (La Salle, Ill: Open Court 1989). Reprinted in HLAE, pp. 109–148.

Diamond, Cora (1966). ‘Secondary Sense’, Proceedings of the Aristotelian Society 67.
Reprinted in The Realistic Spirit, pp. 225–242.

Vogler, Candace (2002). Reasonably Vicious (Cambridge, Mass.: Harvard University
Press).

Von Wright, G. H. (1972). ‘On So-called Practical Inference’, Acta Sociologica, 15/1.

6
THE CHARACTER OF THE
CONCEPT OF INTENTION

(§§42–49)

§§5–19 took us from simplicity to complexity and apparent
disorder, and §§20–40 from complexity to order. In §§41–49
Anscombe seeks to show how our language masks that complexity
in a way that leads us toward the ‘natural’ pictures that she has
been seeking to undermine. We do not notice the complexity
because many of our concepts incorporate the order in their sense.
Our concepts parcel up what is ‘enormously complicated’ into a
form such that ‘a child can give such a report!’ (§43, p. 80). These
sections return us to the simplicity from which we began, but this
time with any inclination we may have had to say that things are
‘intuitively clear’ (§2, p. 1) thoroughly subdued.

Aside from a single sentence in §48 – ‘practical knowledge is
“the cause of what it understands”’ (p. 87) – these paragraphs have
not attracted much attention from commentators. This may be
because reading Intention as an answer to Anscombe’s Question
renders these sections irrelevant and doubly obscure. This neglect
is unfortunate because they contain some of the most profound –
and correlatively most difficult – insights of Intention.

THE CHARACTER OF THE CONCEPT OF INTENTION152

In these sections Anscombe gives her account of the character of
the concept of intention – owing from §1 – along with her final reso-
lution of the puzzle of §28, available now that those two assumptions
about practical reasoning have been dislodged. She also argues –
or, rather, says – that descriptions of human actions, descriptions
like ‘opening’, ‘pumping’, ‘resting’, are descriptions which them-
selves depend for their meaning on the concept of intention (§47)
and, as such ‘the description of something as a human action could
not occur prior to the existence of the question “Why?”’ (§46,
p. 83). These sections turn on its head any philosophical enquiry
which assumes it is possible to identify and describe those events
which are actions prior to giving an account of intention.

FORM OF DESCRIPTION OF EVENTS (§§46–48)

In §47, Anscombe answers the question with which Intention
began. The question was not Anscombe’s Question – that is, it
was not: how do we connect the phenomena referred to in the
three uses of ‘intention’? It was rather: what is the character of
the concept of intention? Anscombe’s answer to that question also
contains a diagnosis of why Anscombe’s Question seemed to arise:

If one simply attends to the fact that many actions can be either inten-
tional or unintentional, it can be quite natural to think that events which
are characterisable as intentional are a certain natural class, ‘inten-
tional’ being an extra property which a philosopher must try to describe.

In fact the term ‘intentional’ has reference to a form of description
of events. What is essential to this form is displayed by the results of
our enquiries into the question ‘Why?’

(§47, pp. 84–85)

I will discuss the second paragraph of this quotation first.
To speak of the ‘form of description of intentional actions’ is

to speak of that order – the order revealed by the question ‘Why?’
and by practical reasoning when it is set out in full. Anscombe says
that ‘[e]vents are typically described in this form when “in order
to” or “because” is attached to their descriptions’ (pp. 84–85).
If I say ‘I slid on the ice because I felt cheerful’ or ‘He operated the

THE CHARACTER OF THE CONCEPT OF INTENTION 153

pump in order to replenish the water supply’ then the description
of what is happening appears in this form. If a description occurs
in this form without the ‘in order to’ or ‘because’ attached, then a
request can be made: ‘Why?’.

Anscombe notes that as well as speaking of ‘the form of descrip-
tion “intentional actions”’ we can also speak of ‘descriptions that
can occur in this form’ (§47, p. 85). We focus on the order when
we use place-holder letters for descriptions of actions – as in her
‘vague and general formula’: ‘I do X in order to do Y’ – but for
a lot of Intention Anscombe uses particular descriptions: ‘sliding
on ice’, ‘operating the pump’, ‘replenishing the water supply’ are
descriptions that occur in this form in the examples we just gave.
Intention is full of others. A mark of a description occurring in this
form is, of course, that the question ‘Why?’ has application.

Clearly not every description can occur in this form – not even
every description that can be true of a person. ‘Turning to stone’
and ‘digesting the contents of her stomach’ cannot. These are not
descriptions of possible intentional actions for a human. But the
range of descriptions that can is very large.

Many descriptions can occur in the form but generally don’t,
while others seem to be made for this form, almost always appear-
ing in it. This can be seen by contrasting ‘sliding’ and ‘operating’
above. The verb ‘to operate’, as it is usually meant, describes an
action, without need for the ‘in order to’ or ‘because’ attachment
to make that clear. But the same is not true of ‘sliding’, which is
perhaps more commonly found outside that form, as in ‘He broke
his leg sliding on the ice’. If I say ‘He is operating the pump’, you
will naturally assume that he is operating the pump in order to ….
You will naturally assume that the question ‘Why?’ has applica-
tion. If I say ‘He is sliding on the ice’, the description invites the
question ‘Intentionally or unintentionally?’.

Anscombe makes a remark which draws our attention to this
contrast on the very first page of Intention:

‘I am going to be sick’ is usually a prediction; ‘I am going to take a
walk’ usually an expression of intention. The distinction intended is
intuitively clear, in the following sense: if I say ‘I am going to fail in this
exam’ and someone says ‘Surely you aren’t as bad at the subject as

THE CHARACTER OF THE CONCEPT OF INTENTION154

that’, I may make my meaning clear by explaining that I was express-
ing an intention, not giving an estimate of my chances.

(§2, pp. 1–2)

The need to clarify – to ‘make my meaning clear’ – occurs here
because ‘failing in this exam’ does not usually occur in this form,
though as the imagined exchange makes clear, it can. ‘I am failing
my exam in order to …’ is sometimes intelligible. The same is true
of the description ‘being sick’: ‘I am going to make myself sick in
order to get this poison out of me’. Later, Anscombe gives another
example in which descriptions that do not generally appear in this
form, do so:

[T]he description in ‘Why are you contracting those muscles?’ is ruled
out if the only sort of answer to the question ‘Why?’ displays that the
man’s knowledge, if any, that he was contracting those muscles is
an inference from his knowledge of anatomy. And the description in
the question ‘Why are you generating those substances in your nerve
fibres?’ will in fact always be ruled out on these lines unless we sup-
pose that the man has a plan of producing these substances (if it were
possible, we might suppose he wanted to collect some).

(§23, p. 38)

These examples show that when a description that doesn’t usually
appear in this form does so, we will need some ‘filling out’. This
filling out will need to make sense both of this person wanting to,
e.g. be sick or generate some substances in his nerve fibres (to get
out of work; as part of a plan) and of the agent thinking ‘I am sick
/ generating those nerve fibres / failing this exam with a view to
Q’ (see §22, p. 36). For these descriptions, the answer ‘No reason’
will often be strange or unintelligible (see §§17–18).

As we saw in Chapter 5, the question of the intelligibility of a
person is connected to our ‘knowledge and opinion’ about what
is the case and what can happen. ‘Postponing an eclipse’ is not a
description that we will recognise as occurring in that form, though
Dean Swift’s parishioners did. So the set of descriptions that can
occur in this form will wax and wane along with our knowledge
and understanding of the world and the sort of technological,

THE CHARACTER OF THE CONCEPT OF INTENTION 155

scientific, and other advances that extend our knowledge of ‘hypo-
theticals’ and our capacity to make them true. ‘Going to the moon’,
‘proving Fermat’s last theorem’ and ‘implanting an embryo’ are all
descriptions that can now appear in this form, though they couldn’t
when Anscombe wrote Intention.1

The sort of ‘filling out’ just described usually isn’t necessary for
descriptions that are ‘made’ for the ‘in order to’ form. If I say ‘I am
going to take a walk’ it would be baffling if you mistook this for
a prediction rather than an expression of intention; ‘I’m making a
cup of tea’ is perfectly intelligible on its own. For such descriptions,
‘No particular reason’ or ‘I just am’ are also often perfectly intelli-
gible. Anscombe does not think that this is just because it is more
common that going for a walk or making a cup of tea is an inten-
tional action while being sick is not. It is not, that is, just a matter
of empirical regularity. Rather, she says that for these descriptions,
it is not just that they can occur in the form, rather that they
are ‘dependent on the existence of this form for their own sense’
(§47, p. 85). If this is right, then these descriptions, dependent as
they are on the order Anscombe has set out, could not exist outside
the linguistic practice characterised by the question ‘Why?’

Anscombe says that this class of descriptions – the class that
depend for their sense on the concept of intention – ‘is a very
large, and the most important, section of those descriptions of
things effected by the movements of human beings which go to
make up the history of a human being’s day or life’ (§47, p. 85).
She identifies two sorts of description within this class:

Intruding
Offending
Coming to possess
Kicking (and other descriptions connoting
characteristically animal movement)
Abandoning, leaving alone
Dropping (transitive), holding, picking up
Switching (on, off)
Placing, arranging

Telephoning
Calling
Groping
Crouching
Greeting
Signing, signalling
Paying, selling, buying
Hiring, dismissing
Sending for
Marrying, contracting

THE CHARACTER OF THE CONCEPT OF INTENTION156

Focusing first on the left-hand list, Anscombe says it contains
‘descriptions which go beyond physics: one might call them vital
descriptions’ (§47, p. 86). These descriptions ‘go beyond physics’
because they are descriptions which evoke the formal order which
is connected with intentional action. Anscombe acknowledges
that the members of the list will ‘strike anyone as a very mixed set’
(§47, p. 85), but what unifies them is that they are ‘all basically at
least animal’, where:

‘characteristically animal movements’ are movements with a normal
role in the sensitive, and therefore appetitive, life of animals.

(§47, p. 86)

This is a difficult point because, as Anscombe acknowledges, she
has ‘defined intentional action in terms of language – the special
question “Why?”’ but is now ‘introducing intention-dependent
concepts with special reference to their application to animals,
which have no language’ (§47, p. 86).

If one is concerned about this that may be a symptom of try-
ing to take the order Anscombe revealed as ‘describ[ing] actual
mental processes’ rather than the criteria for the application
of the concept of intention. The order Anscombe identifies is
meant to characterise a particular use of descriptions, one on
which ‘we describe what further [someone] is doing in doing
something (the latter description being more immediate, nearer
to the merely physical)’ (§47, p. 86). When I say ‘Smith is going
for a walk in order to clear his head’ I describe Smith’s action
in terms of this calculative order, without this implying that
Smith had the thought ‘I will go for a walk in order to clear my
head’. He might have such a thought and yet it be untrue that
he is going for a walk to clear one’s head – suppose he is trying
to convince himself that that is his objective, when really he is
seeking an opportunity to spy on his girlfriend of whom he is
suspicious (see §25, p. 42).

If the concept describes a calculative order and not a process of
calculation then, says Anscombe, we can see that, although animals
do not express their intentions:

THE CHARACTER OF THE CONCEPT OF INTENTION 157

we describe what they do in a manner perfectly characteristic of the
use of intention concepts: we describe what further they are doing in
doing something (the latter description being more immediate, nearer
to the merely physical): the cat is stalking the bird in crouching and
slinking along with its eye fixed on the bird and whiskers twitching.

(§47, p. 86)

The descriptions on the right-hand list also ‘go beyond physics’,
but are additionally ‘descriptions [where] intention is required
(as an extra feature) by the definitions of the concepts employed’
(§48, p. 87).2 Examples like ‘entering into a contract’ and ‘getting
married’ and ‘making a promise’ are cases of this kind. With all
these descriptions ‘it is essential … that someone who is doing it
should think he is doing it’ (1969a, p. 10). If, for example, it was
discovered during a wedding that ‘one of the parties did not know
the nature of the ceremony … that would invalidate a marriage’
(§48, p. 87). Given, then, that these descriptions contain the con-
cept ‘intention’ in their definition, to say that they are dependent
on that concept for their sense is uncontroversial.3

If all this is right then it follows that the ‘quite natural’ thought
that ‘events which are characterisable as intentional are a certain
natural class, “intentional” being an extra property which a phi-
losopher must try to describe’ (§47, p. 84), is utterly mistaken.
On that way of thinking, our ‘special interest in human actions’
is like a special interest in fossils, or stamps, or bacteria, as if
‘we have a special interest in the movement of these molecules –
namely, the ones in a human being; or … in the movements of
certain bodies – namely human ones’ (§46, p. 83). If we think
that way, then we face an additional question: but why are these
movements subject to the question why? And this question may
lead us to look for some property of those movements which
explains its application. This, says Anscombe, would be like
thinking that ‘for some undiscovered reason … certain appear-
ances of chalk on a blackboard are subject to the question “What
does it say?”’. Rather, just as ‘the description of something as
a word or a sentence at all could not occur prior to the fact
that words or sentences have meaning[, s]o the description of

THE CHARACTER OF THE CONCEPT OF INTENTION158

something as a human action could not occur prior to the exist-
ence of the question ‘Why?’ (§46, p. 83).

We can connect this thought with the false view of practical
reasoning, mocked by Anscombe when she spoke of the ‘mince
pie syllogism’. Unlike mince pies, the subject matter of practical
reasoning – human action – would not exist for us to be interested
in, were it not for the fact that we reason practically. Looking back
to our discussion of ‘The Causation of Action’, to think that way
would be to suppose that our interest in ‘how a human being
works’ (p. 93) preceded our interest in human action. Echoing
that argument, Anscombe says now:

Why do we say that the movement of the pump handle up and down
is part of a process whereby those people cease to move about? It is
part of a causal chain which ends with that household’s getting poi-
soned. But then so is some turn of a wheel of a train by which one of
the inhabitants travelled to the house. Why has the movement of the
pump handle a more important position than the turn of the wheel?
It is because it plays a part in the way a certain poisonous substance
gets into human organism, and that a poisonous substance gets into
human organism is the form of description of what happens which here
interests us; and only because it interests us would we even consider
reflecting on the role of the wheel’s turn in carrying the man to his fate.

(§46, pp. 83–84)

In her view, we are interested in how a human being works because
we are already interested in human action. What directs our atten-
tion to certain changes and movements – in humans and in animals,
and in the environment – is the fact that those changes and move-
ments are pumpings, poisonings, killings, buildings, greetings.

DIAGNOSING THE CARTESIAN IMPULSE

Let’s return to the quotation with which we began this section.
Anscombe says:

If one simply attends to the fact that many actions can be either inten-
tional or unintentional, it can be quite natural to think that events

THE CHARACTER OF THE CONCEPT OF INTENTION 159

which are characterisable as intentional are a certain natural class,
‘intentional’ being an extra property which a philosopher must try to
describe.

(§47, p. 84)

In his critical notice of Intention, Kurt Baier connects this ‘natural’
thought with the view that ‘if we want to know a man’s intentions
it is into the contents of his mind, and only into these that we
must enquire’ (§4, p. 9). His insight is that Intention reveals the
way in which the ‘intuitive’ but false picture of the character of the
concept of intention – a picture of its grammar or logical form –
leads to an interiorising theory about the nature of intentions. It
does so when combined with ‘[c]ertain seemingly indubitable
“facts”’ (Baier, 1960, p. 71) – facts such as those described by
Anscombe in §4:

[I]n general we are interested, not just in a man’s intention of doing
what he does, but in his intention in doing it, and this can very often
not be seen from seeing what he does. [Also] in general the question
of whether he intends to do what he does just does not arise (because
the answer is obvious); while if it does arise, it is rather often set-
tled by asking him. And, finally, a man can form an intention which
he then does nothing to carry out, either because he is prevented or
because he changes his mind: but the intention itself can be complete,
although it remains a purely interior thing.

(§4, p. 9)

It is these facts in the context of this distorting grammar, and not
these facts as such, which lead to the difficulty.

Baier’s analysis points to two ways of rejecting the interio-
rising thesis about intention. First, one might say: ‘intention’
refers to something whose existence is not purely in the sphere of the
mind. This would lead one into a debate about whether inten-
tions are ‘inner’ or ‘outer’ or, equivalently, about the boundaries
of the ‘inner’ or of ‘self-knowledge’. It would be a debate about
the nature of intentions and intentional actions, and about pre-
cisely what property ‘intention’ picked out. One might make this
kind of case for rejecting the interiorising account of intention

THE CHARACTER OF THE CONCEPT OF INTENTION160

by showing that intention in action was the most basic case of
intention; or by denying the possibility of pure intending; or by
holding that the mental state of intending stands in need of out-
ward criteria; or (if this is different) by demonstrating that there
was an ‘internal relation’ between an intention to act and an inten-
tion in action. This way of proceeding would leave intact the idea
that ‘intention’ refers to a property that some actions have and
some lack, while rejecting a thesis about the kind of property it
is. When Intention is read as an answer to Anscombe’s Question
this is precisely the sort of debate Anscombe is taken to be having.
But, as Davidson shows, once one concedes to this view of the
character of the concept, the battle is lost already. If one acknowl-
edges the fact that I can have an intention which I never manifest,
then ‘there is no reason not to allow that intention of exactly the
same kind is also present when the intended action eventuates’
(Davidson, 1978, p. 89); or one denies this ‘indubitable fact’ in
favour of an implausible behaviourism.

The alternative, then, is to look at that frame: the assumption
about the character of the concept within which this debate takes
place. Here one contradicts the interiorising claim by saying:
‘intention’ does not refer to some thing or class of things; the concept
does not have that kind of character. According to this diagnosis of
the error, everyone involved in the first debate – the one about the
nature of intentions – has made the very same mistake. They have
left the distinction made by the concept of intention, as something
that is ‘intuitively obvious’; and this is something that Intention
seeks to show cannot be done. It is Anscombe’s challenge to the
‘intuitive’ picture of the character of the concept of intention, and
not to the intuitive account of intentions, that sets her apart from
most contemporary – indeed most Modern – philosophy of action.

When Anscombe says that ‘the term “intentional” has reference
to a form of description of events’ what she means, then, is that
when ‘a man says “I am going to do such and such”’ and we call
that an ‘expression of intention’, we mean: the description ‘doing
such and such’ occurs in this form. When we ‘speak of an action as
intentional’ we mean: the description of what he is doing occurs in
this form. When we ‘ask with what intention the thing was done’
we ask for the ‘because’ or ‘in order to’ attachment which is implied

THE CHARACTER OF THE CONCEPT OF INTENTION 161

when a description of what was done occurs in this form. The con-
cept of intention applies in each case to the description and assigns
it to a calculative order; it does not apply to some state or prop-
erty, mental or physical, of a human being.

In the paragraph quoted above, Anscombe connects the grip of
these false pictures with our natural interest in cases in which the
question ‘intentional or unintentional?’ arises. She says that when
we focus on such cases, and on ‘small sections of action and slips
which can occur in them’, then it will start to seem that inten-
tion is ‘a mere extra feature of events whose description would be
otherwise the same’ (§48, p. 88). So the intuitive picture of the
grammar is linked to our interest in a limited number of examples.
Furthermore, taking these examples, they are those in which it
is most likely to be the case that we discover the answer – to the
question ‘Intention or unintentional?’ by asking the agent. So,
such cases emphasise the interiority or intention. It is when these
two combine that the Cartesian model starts to look irresistible.

As a prophylactic to this view we should look at cases in which
the descriptions themselves are dependent for their sense on the
form of description ‘intentional action’, and cases about which the
question ‘intentional or unintentional’ cannot intelligibly arise.
That, of course, is exactly what Anscombe does in Intention.

PRACTICAL KNOWLEDGE (§§28–32, §§45–48)

As mentioned at the start of this chapter, the sentence in these
late paragraphs that has – almost uniquely – attracted attention is
the following: ‘Practical knowledge is the cause of what it under-
stands, unlike “speculative” knowledge, which is “derived from
the objects known”’ (§48, p. 87). There has been a tendency in
the literature to try to construct an account of practical knowl-
edge from just this sentence and selected remarks from §§28–32
(sections discussed under ‘knowledge without observation’ in the
previous chapter). This tendency misconstrues the status of this
sentence – to which we will come shortly – and obscures the role
of Anscombe’s discussion of practical reasoning in securing her
description of the character of the concept of intention.

THE CHARACTER OF THE CONCEPT OF INTENTION162

In Chapter 5 we saw that the interlocutor objects to Anscombe’s
description of the application of the question ‘Why?’. The descrip-
tion cannot be right, he says, because according to it a person
knows what he is doing without observation. But a person cannot
know what he is doing without observation if what he is doing is
what is happening. Anscombe’s reply, I suggested, has the form:
yes, he can, and you will see this if you understand the character
of practical reasoning.

All three of the voices in Intention – descriptive, intuitive, and
philosophical – are to be heard in that exchange. Descriptive: An
answer to the question ‘Why are you X-ing?’ which reveals that
the person doesn’t know he is X-ing, or knows it only because he
notices it, shows that ‘X-ing’ is not a description under which he
is doing what he is doing. Intuitive: But, you can’t know what is
happening without observation. Philosophical: Here is how.

In the previous chapter we set out Anscombe’s account of
practical reasoning, according to which the conclusion of a piece
of practical reasoning is ‘an action whose point is shewn by the
premises’ (§33, p. 60). I framed that discussion as a response to
the interlocutor’s objection and said that Anscombe identifies
two, connected, features of our ‘natural’ thinking about practical
reasoning, features which stand in the way of the idea that the
conclusion of practical reasoning is an action under a description.
First, the view that practical reasoning is ‘reasoning toward the
truth of a proposition which is supposedly shown to be true by
the premises’ (§33, p. 58). Second, the view that the action that
arises is compelled or necessitated by that reasoning. Now we need
to make explicit how that way of thinking about practical rea-
soning is supposed to give rise to an objection to speaking of a
person knowing what she is doing without observation. I want
to suggest that the first feature leads to the idea that knowledge
of what I am doing ‘must be something that is judged as such by
being in accordance with the facts’ (§32, p. 57) and hence leads to
a characterisation of practical knowledge – the knowledge a man
has of his intentional actions – as distinctive only in terms of its
subject matter. The second feature – the idea that the action that
arises is compelled or necessitated by the reasoning – character-
ises the relation between the reasoning and the action that is its

THE CHARACTER OF THE CONCEPT OF INTENTION 163

conclusion, such that the former is merely ‘a necessary condition
of the production of various results’ (§48, p. 88). This occludes
the fact that the description under which one acts occurs in the
form of description ‘intentional actions’, a form that is displayed
in practical reasoning (§47, p. 84). We now need to bring these
insights to bear on the interlocutor’s objection.

We have already discussed (in Chapter 5) several paragraphs
between §28, where the objection is raised, and §33, where
Anscombe turns to her discussion of practical reasoning. I want
to go back to the final two paragraphs of those sections, in which
Anscombe compares orders and present-tense expressions of inten-
tion. She says that the comparison ‘illuminates the periphery of
the problem’ but ‘fails at the centre and leaves that in … darkness’
(§31, p. 55). We will look first at what the discussion ‘illuminates’.

Anscombe describes how, with both orders and expressions of
intention, there are two, quite distinct, ways in which an error can
lead to ‘a discrepancy between the language and that of which the
language is a description’ (§32, p. 57). One discrepancy ‘imputes
a fault to language’ and the other ‘does not impute a fault to lan-
guage but to the event’ (§32, p. 57).

It is important to note that Anscombe is not interested, here, in
cases that involve ‘ignoring, disregarding, or disobeying an order’
(§32, p. 57). In such cases the would-be order-follower decides
not to follow the order and so there is no error; just a decision not
to act in accord with the imperative. These cases, for reasons that
will become clear shortly, have no parallel in intentional action,
though they do, as we see in Chapter 7, have a parallel in expres-
sions of intention for the future.

To illustrate the first kind of case, in which a discrepancy ‘imputes
a fault to language’, Anscombe gives the following example:

A certain soldier was court-martialled (or something of that sort) for
insubordinate behaviour. He had, it seems, been ‘abusive’ at his med-
ical examination. The examining doctor had told him to clench his
teeth; whereupon he took them out, handed them to the doctor and
said ‘You clench them’.

(§31, p. 55)

THE CHARACTER OF THE CONCEPT OF INTENTION164

Anscombe says that in the face of the fact that this man has no
teeth, ‘the order falls to the ground’ (§32, p. 57). Compare ‘Tie
your shoelaces!’ to a person wearing slip-on shoes, ‘Finish your
homework!’ to a child who has no homework, ‘Supply me with
potatoes!’ to the postman. Because this man has no teeth there is
no possibility of the order being executed, no such thing as obeying
it nor as disobeying it.4 As such the order is at fault. The order-
giver was wrong about the facts – had false opinion about what was
the case and what could happen. His order was not an assertion
that these conditions held, but without their holding his statement
could not fulfil its role as an order. This is why the charge of insub-
ordination is correctly related to the soldier’s ‘abusive’ behaviour,
rather than any sort of disobedience.

In the second kind of case, the discrepancy imputes a fault ‘to
the event’. These are cases of obeying an order wrong:

we ought to be struck by the fact that there is such a thing, and that
it is not the same as ignoring, disregarding, or disobeying an order.
If the order is given ‘Left turn!’ and the man turns right, there can
be clear signs that this was not an act of disobedience. But there is
a discrepancy between the language and that of which the language
is a description. But the discrepancy does not impute a fault to the
language – but to the event.

(§32, p. 57)

We ‘ought to be struck by it’, I suppose, because it is highly
unusual. The example – ‘Turn left!’ and you turn right – is a
fairly familiar one, at least from slapstick comedy, which perhaps
makes this sort of error seem more prosaic. But the reasons it is
familiar are specific to the case: some people are bad at telling
their left from their right; we are imagining cases in which the
order must be obeyed instantly; what the order requires is not
extended over time but a one-off simple action. If we remove
these local factors – make the order ‘Tie your shoelaces!’ or
‘Pump this pump!’ or ‘Open the door!’ – the oddity of this sort
of error becomes more apparent. The ‘more immediate’ the
description – that is, the closer it is to a description of bodily
movement – the rarer the error becomes (§48, p. 87). I find it

THE CHARACTER OF THE CONCEPT OF INTENTION 165

almost impossible to imagine an error of this kind occurring in
response to the order ‘Clench your teeth!’

In this context Anscombe asks us to consider the role played by a
man’s knowledge of what is the case and what can happen when
he, e.g., gives the order ‘Clench your teeth!’. This is to put in its
proper place the role played by observation in relation to orders.

Think about the first kind of case – the one where the order
‘falls to the ground’. The fault attaches to the order in virtue of
the fact that the doctor was mistaken about what was the case
and what could happen. He thought this soldier had teeth; he did
not. Once he finds out, by observation, that the soldier has no
teeth, he will – and must if he is to remain intelligible – refrain
from giving the order. If he had been giving the order with some
particular end in view, for example, if his aim had been to test the
soldier’s obedience, he may give a different order instead: ‘Touch
your toes!’ If it had been to test the strength of his teeth or to
obey the order given to him by the Sergeant – ‘Have all the men
clench their teeth!’ – then he would have realised that that was not
a possible aim given the situation. Note that in the second case,
where the doctor’s order is in response to an order given to him,
the absence of this man’s teeth invalidates the Sergeant’s order too.

Given this relation between facts and orders, it is possible to
negate, criticise or reject orders by pointing out things about the
situation that would have to be different were the order not to
‘fall to the ground’. To do so is not to contradict the order – it is to
show that no order has been given. The contradictory of the order
‘Clench your teeth!’ is not ‘This man has no teeth’ but ‘Do not
clench your teeth!’ (§31, p. 55).

What about the second case, where the mistake is ‘in the event’?
Note first that this second kind of error excludes the first kind; if
an order ‘falls to the ground’ then there is no such thing as follow-
ing it at all – nor indeed as ignoring, disregarding or disobeying it.
There is, in a sense, no order given, just an attempt made to give
an order. Given that there is no such thing as the second kind of
mistake unless the order has, in fact, been given – that is, if the
order ‘falls to the ground’ then there is no such thing as obey-
ing the order wrong – the possibility of obeying the order wrong

THE CHARACTER OF THE CONCEPT OF INTENTION166

depends on the facts being as they need to be for the order to be
given. So, there is a dependence on the facts in that sense. But,
crucially, if the order does not fall to the ground then the ques-
tion of correctness or error consists solely in what the would-be
order-follower does. This kind of error, then, does not involve a
mistake about what is the case.

Considering orders is a good way to get clear about the distinction
between these two kinds of error, and the role played by obser-
vational knowledge in relation to each. Anscombe argues that
the very same relations between what is said, the facts and what is
done hold in the case of expressions of intention has in the case of
commands. As we will see, these relations are occluded by certain
features of the ‘superficial grammar’ (§2, p. 4) of expressions of
intention.

To outline the case for the parallelism, Anscombe describes
what has become a famous case of a man out shopping:

Let us consider a man going round town with a shopping list in his
hand. Now it is clear that the relation of this list to the things he actu-
ally buys is one and the same whether his wife gave him the list or it is
his own list; and that there is a different relation when a list is made by
a detective following him about. If he made the list himself, it was an
expression of intention; if his wife gave it to him, it has the role of an
order. What then is the identical relation to what happens, in the order
and in the intention, which is not shared by the record? It is precisely
this: if the list and the things that the man actually buys do not agree,
and if this and this alone constitutes a mistake, then the mistake is not
in the list but in the man’s performance (if his wife were to say: ‘Look,
it says butter and you have bought margarine’, he would hardly reply:
‘What a mistake! we must put that right’ and alter the word on the list
to ‘margarine’); whereas if the detective’s record and what the man
actually buys do not agree, then the mistake is in the record.

(§32, p. 56)

Anscombe points out (descriptive voice) that the question of
whether the man’s shopping list is an order or an expression of
intention turns only on whether ‘his wife gave him the list’ (order)

THE CHARACTER OF THE CONCEPT OF INTENTION 167

or ‘he made the list himself ’ (expression of intention). Either way,
says Anscombe, ‘the relation of this list to the things he actually
buys is one and the same’. What he buys must match what is on
the list if he is not to be in error (or disobeying, changing his
mind, or suffering weakness of will).

This example is helpful for two reasons. First, it allows
Anscombe to remove a feature of ‘superficial grammar’ that
might otherwise obscure the parallel. Spoken orders and spoken
expressions of intention tend to make use of different sentential
vehicles. An order is usually cast in the imperative mood: ‘Do X!’.
An expression of intention in the indicative: ‘I will do X’, ‘I am
doing X’. (That this is not always so is illustrated by Anscombe
in §2; ‘Nurse will take you to the operating theatre’, said by a
doctor to a patient in the presence of a nurse, ‘may function
both as an expression of his intention … and as an order’ (p. 3).)
By moving to a list ‘Bread, Margarine, Milk’, this difference in
sentential vehicle disappears. Second, by making the action one
which involves actual objects – bread, margarine, milk – there can
be no temptation to seek agreement with the order or intention
in something internal.

The advantage of introducing the parallel with orders is that
the distinction between the two kinds of error is much clearer
there than it is when we look at expression of intention. There
are two reasons for this. First, with orders one kind of discrep-
ancy leads to criticism of the would-be order-giver and the other
to criticism of the would-be order-follower. We say in the first
case, to the doctor: ‘You shouldn’t have ordered the soldier to
clench his teeth – he doesn’t have any’; in the second we say to
the solider: ‘The Sergeant told you to turn left, and you turned
right!’. With expressions of intention the agent is both, as it
were, order-giver and order-follower. This makes it easier to elide
the two kinds of mistake because in both the agent himself is
at fault. Instead of imparting fault to different people, we must
impart fault first to the agent’s intention and second to his action.
Second, expressions of intention, unlike orders, are generally –
though as the shopping-trip example shows, not always – cast in
the indicative mood, ‘I am doing such-and-such’. Thus, where
we have been speaking of error in relation to orders in terms of

THE CHARACTER OF THE CONCEPT OF INTENTION168

a ‘discrepancy between the language and that of which the lan-
guage is a description’ (§32, p. 57), we can speak of falsehood
in the case of expressions of intention. When I say ‘I am doing
such-and-such’, either kind of error – in language or in the event –
will make what was said false: ‘doing such-and-such’ will not be
a description of what I am, in fact, doing. As we’ll see, this fact –
that expressions of intention are the kind of statement that can
be true or false – is a source of the ‘natural’ idea that ‘I am doing
such-and-such’ must be judged as knowledge by being in accord-
ance with the facts, and creates ‘a point at which the parallelism
[between orders and description by the agent of his present inten-
tion] ceases’ (§31, p. 55). This is where we will need Anscombe’s
account of practical reasoning.

That is to get ahead of ourselves. Let us begin by setting out
the parallelism.

Anscombe describes a way in which a statement of what one is
doing, ‘I am replenishing the water supply’, can ‘fall to the ground’
in just the same way as an order can. This can happen when ‘some-
thing is not the case which would have to be the case in order for
his statement to be true; as when, unknown to the man pump-
ing, there was a hole in the pipe round the corner’ (§32, p. 56).
Anscombe says that the fact that there is a hole in the pipe, ‘relates
to his statement that he is replenishing the water-supply as does
the fact that the man has no teeth of his own to the order “Clench
your teeth”’ (§32, p. 56).

Note that while ‘Clench your teeth!’ does not have the right
kind of grammatical form to be called false – it is in the impera-
tive mood – ‘I am replenishing the water supply’ does. So where
we speak of an order as ‘falling to the ground’ we can speak of an
expression of intention as false.5 However, what Anscombe wants
the parallel to illuminate, is that what it is for a statement of inten-
tion to be false in this way is quite different from what it is for
an observational statement to be false. Recall again the relation
between the fault in ‘Clench your teeth!’ and the sort of facts that
could be the object of observational knowledge: for example, that
this man doesn’t have any teeth. These facts must be in place if
the order is not to ‘fall to the ground’, but the order is not a claim
or judgement that those facts are in place – it assumes they are.

THE CHARACTER OF THE CONCEPT OF INTENTION 169

Similarly, if our would-be water-replenisher discovers that the pipe
is damaged, he will not carry on pumping. He will recognise that
his intention ‘falls to the ground’ along with every other descrip-
tion in the order set out using ‘Why?’ which was, so to speak,
further along the chain. Were he to carry on, his action – and
he himself – would cease to be intelligible; we would need some
‘filling out’ to make it so.

This illuminates a difference in the relation between the state-
ment ‘p’ and the statement ‘~p’ when the former is an observational
judgement and when it is an expression of intention. Anscombe
uses the parallel with orders to bring this out:

What is the contradictory of a description of one’s own intentional
action? Is it ‘You aren’t, in fact’? – E.g. ‘You aren’t replenishing the
water supply, because water is running out of a hole in the pipe’? I sug-
gest that it is not. […] The statement: ‘The water is running out of a pipe
round the corner’ stands in the same relation to the statement ‘I’m
replenishing the water supply’ as does ‘My teeth are false’ to the order
‘Clench your teeth’; and so the statement (on grounds of observation)
‘You are not replenishing the water-supply’ stands in the same relation
to the description of intentional action ‘I am replenishing the water
supply’, as does the well-founded prediction ‘The man isn’t going to
clench his teeth, since they are false’ to the order ‘Clench your teeth’.

(§31, p. 55)

The crucial point is to not let the fact that ‘You aren’t replenishing
the water supply’ implies that ‘I am replenishing the water supply’
is false, occlude the parallel with orders. Just as ‘You aren’t clench-
ing your teeth’ is not a contradiction of ‘Clench your teeth!’, so
too a description of the facts – ‘You aren’t replenishing the water
supply’ – cannot be a contraction of an expression of intention.

At this point it will be instructive to remind ourselves of
Anscombe’s first reply to the interlocutor’s objection:

My reply is that the topic of an intention may be a matter on which
there is knowledge or opinion based on observation, inference, hear-
say, superstition or anything else that knowledge or opinion are ever

THE CHARACTER OF THE CONCEPT OF INTENTION170

based on; or again matter on which an opinion is held without any
foundation at all. When knowledge or opinion are present concerning
what can be the case, and what can happen – say Z – if one does cer-
tain things, say ABC, then it is possible to have the intention of doing
Z in doing ABC; and if the case is one of knowledge or if the opinion is
correct, then doing or causing Z is an intentional action, and it is not
by observation that one knows one is doing Z.

(§28, p. 50; emphasis added)

We can now see the role of the phrase to which I have added
emphasis. Our would-be poisoner has an opinion about what is
the case – here is a pump, the pump is connected to a pipe, the
pipe leads to the house, the water in the tank has been poisoned –
and about what can happen if he moves his arm up and down
with his hand on the pump handle: thus ‘it is possible [for him]
to have the intention of poisoning the inhabitants of the house
in moving his arm up and down’. But if the opinion is incorrect,
then ‘poisoning the inhabitants’ is not a description of his inten-
tional action, no matter what he thinks. And here, the fault is not
with what he does but with his intention. His practical reasoning
was at fault.

We now come to the second kind of error. Recall that a con-
dition for this kind of error is that the first kind of error has
not occurred. So, we must suppose that the practical reasoning
was sound and valid and so our agent has knowledge, or at least
true opinion, of what is the case and what can happen, and that
he has drawn his conclusion in action. In the passage above,
Anscombe says that if these conditions are fulfilled ‘then doing
or causing Z is an intentional action’ and ‘it is not by observation
that one knows one is doing Z’. So, it looks like she wants to
say that so long as intention does not ‘fall to the ground’, what
happens is Z-ing and the agent knows it without observation. In
such a case, when the agent says ‘I am Z-ing’, ‘Z-ing’ will give
the description under which he acts and will be a description of
what he is doing. Why is this not an adequate response to the
interlocutor? Why not say now: he knows what he is doing with-
out observation?

THE CHARACTER OF THE CONCEPT OF INTENTION 171

The difficulty arises with the possibility of a mistake:

[I]s there not possible another case in which a man is simply not doing
what he says? As when I say to myself ‘Now I press Button A’ – pressing
Button B – a thing which can certainly happen. This I will call the direct fal-
sification of what I say. And here, to use Theophrastus’ expression …, the
mistake is not one of judgment but of performance. That is, we do not
say: What you said was a mistake, because it was supposed to describe
what you did and did not describe it, but: What you did was a mistake,
because it was not in accordance with what you said.

It is precisely analogous to obeying an order wrong.
(§32, p. 57)

Recently, in an actual case of this, the socialist leader of the Spanish
opposition party, Pedro Sanchez, pressed the wrong button and
voted in favour of a proposal to restrict access to abortions. We
can imagine him accompanying his action of pressing the ‘Yes’
button with the words ‘I am pressing the “No” button’.

It is the case of this second kind of error that makes it difficult
to see how, when there is no error, we can have a case of knowl-
edge – and here the parallelism with orders ceases to help. In §31
Anscombe says that the comparison ‘illuminates the periphery’ of
the problem of §28, but ‘fails at the centre’. It fails because ‘we say
that the agent’s description [of what he is doing] is knowledge,
but an order is not a piece of knowledge’ (§31, p. 55). The parallel
with orders runs into the ground ‘just where we begin to speak of
knowledge’ (§31, p. 55).

It is at this point that Anscombe makes her now famous appeal
to the concept of practical knowledge, something she says can
‘only be understood if we first understand “practical reasoning”’
(§32, p. 57).

Can it be that there is something that modern philosophy has blankly
misunderstood; namely what ancient and medieval philosophers
meant by practical knowledge? Certainly in modern philosophy we have
an incorrigibly contemplative conception of knowledge. Knowledge
must be something that is judged as such by being in accordance with

THE CHARACTER OF THE CONCEPT OF INTENTION172

the facts. The facts, reality, are prior and dictate what is to be said, if
it is knowledge. And this is the explanation of the utter darkness in
which we have found ourselves.

(§32, p. 57)

Having described practical reasoning as reasoning the conclusion
of which is an action under a description, Anscombe returns to
the case ‘precisely analogous to obeying an order wrong’, now with
the concept of practical knowledge available to her. In a much-
discussed – and much-disputed example – she says:

I wrote ‘I am a fool’ on the blackboard with my eyes shut. Now when
I said what I wrote, ought I to have said: this is what I am writing, if
my intention is getting executed; instead of simply: this is what I am
writing?

Orders, however, can be disobeyed, and intentions fail to get exe-
cuted. That intention for example would not have been executed if
something had gone wrong with the chalk or the surface, so that the
words did not appear. And my knowledge would have been the same
even if this had happened. If then my knowledge is independent of
what actually happens, how can it be knowledge of what does happen?
Someone might say that it is a funny sort of knowledge that was still
knowledge even though what it was knowledge of was not the case!
On the other hand Theophrastus’ remark holds good: ‘the mistake is
in the performance, not in the judgment’.

(§45, p. 82)

I do not want to dwell on the numerous interpretative disputes
that have arisen in relation to this remark, but will comment
briefly on their general flavour, before making my own sug-
gestion. The focus, unsurprisingly, has been what Anscombe
means by saying that ‘my knowledge would have been the
same’ even if she hadn’t written ‘I am a fool’ on the blackboard.
Philosophers who are keen to find in Anscombe’s discussion of
non- observational and practical knowledge the ‘philosophical
treasure’ of an insight into ‘the way in which the mind is embod-
ied’ (Vesey, 1963, p. 210) have been especially distressed by the
seeming suggestion that I can know I am writing ‘I am a fool’

THE CHARACTER OF THE CONCEPT OF INTENTION 173

on the blackboard even when it is false that I am writing ‘I am
a fool’ on the blackboard. This, they have feared, is a first step
toward a bifurcated conception of knowledge of action, itself
a step backwards toward the Cartesian’s interiorised model of
intention. The worry is that if I count as knowing I am doing
such-and-such even when ‘RW is doing such-and-such’ is false,
then the object of my knowledge cannot be what is happening.
Practical knowledge, it seems, must fall short of the facts. In
light of this, many commentators have charged Anscombe with
an error in §45, and argued that what she should have said was
that in such a case the agent lacks practical knowledge.

I want to suggest that this worry – and hence this friendly revi-
sion – is misplaced. Anscombe can say that practical knowledge,
in the case described, is ‘still knowledge even though what it was
knowledge of was not the case’, without this implying that, when
things go well, I cannot have non-observational knowledge of
what I am doing.

We must all agree that when ‘writing “I am a fool” on the
blackboard’ is not a description of what I am doing, then ‘I am
writing “I am a fool” on the blackboard’ is false. If I am not writ-
ing “I am a fool” on the blackboard, then what I say cannot be
true. It is usually assumed that to insist that Anscombe’s writer has
practical knowledge, is to insist that, in some sense, the sentence
is true. I want to suggest that to assume this, and to think correl-
atively that in such a case an agent must lack practical knowledge
of what he is doing, is precisely to affirm what Anscombe denies:
that practical knowledge is a species of speculative knowledge. It
would be to revert to the ‘incorrigibly contemplative conception
of knowledge’ (§32, p. 57).

Consider again the case of knowledge of limb position.
Anscombe says that we have a capacity to say how our limbs are
arranged, but that we can speak of knowing – rather than just being
able to say – because what is said has correctness that are inde-
pendent of our saying it. Such knowledge is speculative, though
non-observational, because it is knowledge ‘which must be judged
as such by being in accordance with the facts’ (§32, p. 57).

When we judge ‘I am writing “I am a fool” on the blackboard’
as failing to be knowledge because it is not in accordance with the

THE CHARACTER OF THE CONCEPT OF INTENTION174

facts, we judge it as a failure of speculative knowledge. However,
we can also ask whether it is a failure of practical knowledge. We
can do so by enquiring as to whether the person who is acting
is properly exercising her capacity for practical reason. If what
she says ‘falls to the ground’, because she is in error either about
what is the case, or about what can happen, then her reasoning
is flawed. However, if she is right about all that, then she can say
what she is doing. What makes it right to speak of knowledge
here, rather than merely a capacity to say – as with ‘giving the
place of one’s pain’ (§8, p. 14) – is that the capacity for practical
reason is itself a capacity contains the possibility of erroneous
exercise. The kind of error that pertains to practical knowledge
is not, however, an error about the facts; rather, it is an error in
practical rationality.

What the comparison with orders is meant to illuminate is
that the match between the description of the intention and the
description of the action – the fact that practical knowledge and
speculative knowledge usually stand and fall together – arises from
the conditions that must be met if the intention is not to ‘fall to
the ground’. These conditions guarantee that, on the whole, when
I have practical knowledge what I say is true. In such cases, I have
non-observational speculative knowledge. However, as we have
seen, it is possible to describe cases in which this is not the case.

Let us end by returning to the sentence that has attracted the most
attention from commentators: ‘practical knowledge is “the cause
of what it understands”’ (§48, p. 87). It is important to see that
this is not a definition or explanation of practical knowledge. If we
look to this sentence for an account of practical knowledge, we
will be disappointed. Rather, as Anscombe says, if we put together
the considerations of §§32–48, we can see what it is for Aquinas’s
account of practical knowledge – that it is the cause of what it
understands – to hold. Without that discussion – in particular,
without recovering Aristotle’s account of practical reasoning –
Aquinas’s meaning could not be understood. Those considera-
tions describe what it is for practical knowledge to be the cause
of what it understands: when a description appears in the form of

THE CHARACTER OF THE CONCEPT OF INTENTION 175

‘ intentional action’, and when there is no error in what is said or
what is done, then it is an agent’s practical knowledge which makes
it the case that what he does ‘come[s] under the description –
execution of intentions’ (§48, p. 88). Anscombe’s Intention is a
rendering of Aquinas’s pre-Modern account for philosophy after
the linguistic turn.

NOTES

1 See Anscombe’s discussion of Wittgenstein and Moore on the proposi-
tion ‘I have never been to the moon’, in her 1976a.

2 Anscombe says that ‘crouching’ may ‘be the only one that occasions any
doubt’. The OED entry for ‘crouching’ makes explicit the connection: ‘To
stoop or bend low with general compression of the body, as in stoop-
ing for shelter, in fear, or in submission; to cower with the limbs bent.
Formerly often applied to the act of bowing low in reverence or defer-
ence. Now said also of the depressed and constrained posture assumed
by a beast in fear or submission, or in order to make a spring.’

3 For a discussion of this important class of concepts, see Anscombe’s
‘On Promising and its Justice, and Whether it Need be Respected in Foro
Interno’ (1969a).

4 Wittgenstein was extremely interested in the relation between orders
which failed in something like this way and something like creativity. The
soldier could for example, have taken his teeth out and clenched them
in his fist; would this have counted as following the order? It seems not.
He describes cases in which the person giving the order does not yet
know what it would be for it to be fulfilled: ‘Do something which I shall
be inclined to call “X”’ is the form he uses. See Cora Diamond (1977).

5 It is interesting to note that between the first and second edition of
Intention, Anscombe makes a slight change to §2 – one of a handful
of alterations. The second edition, that which is standardly referenced,
asks: ‘what are the reasons other than a dispensable usage for not
calling commands true and false according as they are obeyed or dis-
obeyed?’. The first edition makes it clear that this question is meant
rhetorically: ‘in fact there is no reason other than a dispensable usage
why we should not call commands true and false according as they were
obeyed or disobeyed’.

THE CHARACTER OF THE CONCEPT OF INTENTION176

SUGGESTED READING

Anscombe, G. E. M (1965a). ‘Thought and Action in Aristotle: What is “Practical
Truth”?’. From J. R. Banbrough (ed.), New Essays on Plato and Aristotle (London).
Reprinted in FPW, pp. 66–77.

Haddock, Adrian (2011). ‘“The Knowledge That a Man Has of His Intentional
Actions”’. In Anton Ford, Jennifer Hornsby and Frederick Stoutland (eds), Essays on
Anscombe’s Intention (Cambridge, Mass.: Harvard University Press), pp. 147–169.

Schwenkler, John (2015). ‘Understanding “Practical Knowledge”’. In Philosophers’
Imprint, 25(15; June).

Thompson, Michael (2011). ‘Anscombe’s Intention and Practical Knowledge’. In
Anton Ford, Jennifer Hornsby and Frederick Stoutland (eds), Essays on Anscombe’s
Intention (Cambridge, Mass.: Harvard University Press), pp. 198–210.

So far we have limited our discussion to the class of descriptions
of what a man is doing under which he knows he is doing it.
This class also includes answers to the question ‘Why are you
X-ing?’ which are expressions of intention and give one’s further
intention in acting. We have not said anything much on the
topic of expressions of intention for the future – Anscombe’s
first head. As was clear from the last chapter, §§43–49 have
the character of a conclusion, an answer to the question with
which Intention began. Between §§4–49 we move from illusory
simplicity, to chaotic complexity, to clarity, all without saying
much about expressions of intention for the future. It is rather
strange, then, to find §§4–49 bookended by discussion of these
expressions.

Expressions of intention for the future do lie outside the core
of Anscombe’s discussion. However, as we saw in Chapter 5, these
expressions have a very important part to play in understanding
why it is that we take such interest in the concept of intention.

7
EXPRESSIONS OF INTENTION

FOR THE FUTURE
(§§2–3 AND §§50–52)

EXPRESSIONS OF INTENTION FOR THE FUTURE178

THE PLACE OF EXPRESSIONS OF INTENTION
FOR THE FUTURE IN INTENTION

The discussion of expressions of intention for the future is split
between §§2–3 and §§50–52 (the last two sections of the book).
Reading Intention one might be tempted to think that these
sections could be lopped off without much impact on the remain-
ing 44 sections, and indeed the way in which I have presented
things in this Guidebook will have added to this impression.

In §§2–3 Anscombe identifies the importance of reasons
for acting in an account of intention, and this provides a prag-
matic reason for starting with them rather than with intentional
actions. Considering expressions of intention in the future tense,
allows Anscombe to start with a simple contrast between reasons
for thinking true and reasons for acting: a prediction is justified
by reasons for thinking true, an expression of intention for the
future by reasons for acting. As she points out in §22, ‘[t]he ques-
tion of cause versus reasons, which has plagued us in relation to
answers mentioning the past, simply does not arise’ in relation
to future action (p. 34). However, this fact does not make those
sections integral to the later passages in the way, say, that the
discussion of practical reason is integral to the account of inten-
tional action. Indeed, when Anscombe turns to her topic under
its second head in §4, this is something of a fresh start. The final
two paragraphs of Intention, where Anscombe returns to her first
topic, do little more than apply the findings of the previous par-
agraphs. Anscombe says: ‘what I have said about intention in
acting applies also to intention in a proposed action’ (§50, p. 90).
What follows is a brief and focused discussion of the sort of con-
tradiction involved in ‘“I am going to do it” said as an expression
of intention, and “I am not going to do it” as a belief on evidence’
(§52, p. 91–92).

One way to think about this marginalizing of expressions of
intention for the future in Intention would be through the lens
of Anscombe’s Question. The idea would be that Anscombe
affords explanatory priority to intentional actions, hence the pri-
ority she gives them in her book. I want to suggest a different
answer which is connected with our earlier discussion of §20:

EXPRESSIONS OF INTENTION FOR THE FUTURE 179

Anscombe thinks that expressions of intention for the future
are a feature of our linguistic practice but are not essential to
any linguistic practice in which we could speak of intentional
action. Their importance, Anscombe said, is not that without
them there would be ‘no place for the concept of intention at all’
(§20, p. 31); but that without them the concept of intention,
and the description of an action as intentional, would lose ‘a
great deal of its point’ (§20, p. 31). The necessity of these expres-
sions of intention comes, I will suggest, from what is ‘a necessity
for human life’ (Anscombe, 1969a, p. 18) and not from within
the linguistic practice. Anscombe calls these kinds of necessities
‘Aristotelian necessities’: they are things without which ‘good
cannot be attained’ (Anscombe, 1969a, pp. 18–19).

SPECIES OF PREDICTION (§§2–3)

Let us begin by returning to the very start of Anscombe’s enquiry
into the concept of intention. The question she begins with is:
‘[W]hat is the difference between e.g. “I am going to be sick” as
it would most usually be said, and “I am going to take a walk”,
as it would most usually be said?’ (§2, p. 1). We saw in the last
chapter that this remark hints at the point Anscombe reaches
in §48: ‘many of our descriptions of events effected by human
beings are formally descriptions of executed intentions’ (p. 87).
One difference, we can now say, is that ‘taking a walk’ is formally
a description of an executed intention, while ‘being sick’ is not.

In fact, these opening paragraphs contain many of the book’s
central insights. What we find, somewhat surprisingly, is that two
ways of thinking about error – which we saw played a crucial
role in Anscombe’s account of intentional action – are present in
the opening paragraphs. §2 contains core insights into truth and
falsity as it relates to intention, and into the character of practi-
cal knowledge. They connect these insights with ‘Theophrastus’
Principle’ – that there is a kind of error distinctive of intentional
action in which ‘the mistake … is one of performance not judge-
ment’ (§2, p. 5). As we read these opening paragraphs much of
what we find will be familiar.

EXPRESSIONS OF INTENTION FOR THE FUTURE180

Anscombe begins:

The distinction between an expression of intention and a prediction
is generally appealed to as something intuitively clear. ‘I am going to
be sick’ is usually a prediction; ‘I am going to take walk’ usually an
expression of intention. The distinction intended is intuitively clear, in
the following sense: if I say ‘I am going to fail in this exam’ and some-
one says ‘Surely you aren’t as bad at the subject as that’, I may make
my meaning clear by explaining that I was expressing an intention, not
giving an estimate of my chances.

(§2, pp. 1–2)

Anscombe argues that ‘leaving the distinction between estimates
of the future and expressions of intention as something that is
intuitively obvious’ leads one to give an ‘introspective explanation’
(Contents, §2, p. i) which is characterised by ‘various dead-ends’.
Her list of ‘dead-ends’ includes: ‘psychological jargon about
“drives” and “sets”; reduction of intention to a species of desire,
i.e. a kind of emotion; or irreducible intuition of the meaning of
“I intend”’ (§3, pp. 5–6). I want to cast these comments in light
of the distinction between a psychological and a conceptual – or
grammatical, or logical – approach to the topic of Intention.

To leave the distinction between estimates of the future and
expressions of intention as intuitively obvious is to fail to ‘ask […]
what each of these is’, where – for Anscombe at least – to ask this
is to undertake the sort of investigation into the concept of inten-
tion that Intention contains. As we have seen, this investigation
reveals a complicated linguistic practice, embedded in a compli-
cated form of life, and ultimately, if Anscombe is right, then we
come to see that the answer requires us to recognise that many
of our descriptive concepts ‘are dependent on the existence of
[the form of description of “intentional actions”] for their sense’
(§47, p. 85). If we do not carry out this investigation, then we
will be inclined to characterise the difference between these two
uses of language as something simple without noticing that when
we ‘make our meaning clear’, simply by saying ‘I was expressing
my intention’, we are evoking a formal order which exists as part
of this complex form of life. Leaving the distinction intuitively

EXPRESSIONS OF INTENTION FOR THE FUTURE 181

obvious leads us away from ‘considering the verbal expression of
intention’ and toward an explanation in terms of ‘what it is an
expression of ’. Anscombe claims that if we try to say what it is
an expression of we will find ourselves using ‘psychological jargon
about “drives” and “sets”’ and so forth, as we struggle to describe
the properties of such a state.

Anscombe sketches an example:

Suppose it is said: ‘A prediction is a statement about the future’.
This suggests that an expression of intention is not. It is perhaps the
description – or expression – of a present state of mind, a state which
has the properties that characterise it as an intention. Presumably
these are yet to be discovered.

(§2, p. 2)

What sort of properties would such a state have to have? Well,
for one thing, it would need to be ‘essentially connected with
the future’ as ‘no one is likely to believe that it is an accident,
a mere fact of psychology, that those states of mind that are
intentions always have to do with the future’ (§2, p. 2). But the
connection it has to the future would have to be a strange one
because ‘if I do not do what I said I would, I am not supposed
to have made a mistake, or even to have lied; so it seems that
the truth of a statement of intention is not a matter of doing
what I said’ (§2, p. 4). Furthermore, when asked for reasons in
connection with an expression of intention I neither give rea-
sons ‘suggesting what is probable’ nor evidence about my state
of mind but reasons ‘suggesting what it would be good to make
happen’ (§2, p. 4). The task then would be to describe a state
with these properties (cf. §47, p. 84).

In a notoriously difficult argument in §19 Anscombe seeks to
demonstrate, by reductio, that there could be no such property
because the idea that the ‘the content of a description’ with ‘an
internal relation to a description of an action’ could be determined
by looking just at ‘the man by himself in the moment’ is incoher-
ent (§19, pp. 28–9).1 In §2 and §3 she gives two other, absurdly
brief, reductio-type arguments also targeted at the idea that there
could be such a property (§2, p. 2; §3, p. 6).2 Recognising the

EXPRESSIONS OF INTENTION FOR THE FUTURE182

nature of Anscombe’s method, we can see that the conclusion of
these arguments is not, in fact, essential to her enquiry. For sup-
pose that someone could describe a state with those properties, a
state that could play the role that Anscombe describes. This would
not show that there were such a state, only that there being such a
state was not unintelligible. The role of these reductio’s in Intention
is, in that sense, preparatory. By undermining the idea that such
a state could be described, Anscombe means to loosen the grip of
the assumption that the distinction is intuitively obvious, and that
all that needs to be done is to ‘try to describe’ this ‘extra property’
(§47, p. 84). Her aim is to make us less resistant to her suggestion:
that we look away from the phenomena as such, and toward the
linguistic practice.3

Anscombe insists, in §§2–3, that we focus instead on the ver-
bal expression of intention. This leads to a different kind of
‘dead-end’ – an account that is ‘not false but rather mystifying’
(§3, p. 6–7). It is to find a way out of that ‘dead-end’ that she
shifts to an examination of what a man is currently doing.

Anscombe begins her enquiry by asking for the difference
between two uses of the future tense illustrated by ‘I am going to
be sick’ and ‘I am going to take a walk’. The first we call a predic-
tion and the second an expression of intention for the future. She
starts by giving an account of prediction:

The following seems promising: a man says something with one
inflection of the verb in his sentence; later that same thing, only with
a changed inflection of the verb, can be called true (or false) in face of
what has happened later.

(§2, p. 2)

If this was meant to be an account of the difference between
expressions of intention and predictions it clearly fails because,
by this criterion ‘expressions of intentions will also be predictions’
(§2, p. 2). However, that this account of predictions does not
capture the distinction between ‘I am going to be sick’ and ‘I am
going to take a walk’ does not, says Anscombe, imply that it is
wrong:

EXPRESSIONS OF INTENTION FOR THE FUTURE 183

In view of the difficulties described above, this may not consti-
tute an objection. Adopting a hint from Wittgenstein (Philosophical
Investigations, §629–30) we might then first define prediction in
general in some fashion and then, among predictions, distinguish
between commands, expressions of intention, estimates, pure proph-
ecies, etc. The ‘intuitively clear’ distinction we spoke of turns out to be
a distinction between expressions of intention and estimates.

(§2, p. 3)

Chapter 4 highlighted a similar kind of progress made in distin-
guishing reasons from causes. The intuitively clear distinction
turned out to be one between mental causes and reasons associated
with backward-looking motives. These were both answers to the
question ‘Why?’ which mentioned something past, but only in
the former the ‘action has to be thought of by an agent as doing
good or harm of some sort, and the thing in the past as good or
bad’ (§14, p. 22).

Having defined prediction in such a way that expressions
of intention for the future count as a species of prediction,
Anscombe’s task then becomes to ‘distinguish between com-
mands, expressions of intention, estimates, pure prophesies, etc.’
(§2, p. 2). This yields the following account:

an expression of intention is a description of something future in
which the speaker is some sort of agent, which description he justifies
(if he does justify it) by reasons for acting, sc. reasons why it would be
useful or attractive if the description came true, not by evidence that
it is true.

(§3, p. 6)

Anscombe immediately remarks: ‘having got so far, I can see now-
here else to go along this line, and the topic remains mystifying’
(§3, p. 6).

When one first reads Anscombe saying that her account is
‘mystifying’ it is hard not to be baffled. The account she gives here
is just the sort of thing one might expect to find in a philosophy
book; it looks like a perfectly respectable philosophical thesis. The
usual way to criticise such a thesis would be to say either that

EXPRESSIONS OF INTENTION FOR THE FUTURE184

it was false – especially: that there were counter-examples which
revealed the condition was either not necessary or insufficient – or
that the argument in support of it was inconclusive, invalid, or
unsound. To say that it is true but mystifying reveals that some-
thing different is going on here. We now know what.

One reason that the account is ‘mystifying’, of course, is that
it includes the unanalysed notion of a ‘reason for acting’. In
Anscombe’s account, the complexity that a psychological account
would build into the state of intending is relocated to the linguistic
practice within which a description of a possible future state of
affairs can be an expression of intention for the future. And the
formal order that is essential to this practice is best seen by looking
at ‘what kind of true statements about people’s intentions … we
[can] certainly make’ (§4, p. 7).

There is a second reason why the account is ‘mystifying’. So
far I have set out the descriptive part of these opening sections.
But the interlocutor’s voice is also heard here. Central is the fact
that describing ‘I am going for a walk’ is a prediction sounds very
odd. Anscombe notes that ‘it is natural to feel an objection both
to calling commands and to calling expressions of intention,
predictions’ and, she adds, ‘our objections are deep[…] rooted’
(§2, p. 4). In fact, we will see that they have precisely the same
root as the objection of §28.

We call a prediction false when what is said would happen
does not happen. If I predict that Shrewsbury Town F. C. will
win promotion, and they do not, then what I said was false. But
we do not find it natural to speak of expressions of intention in
the same way. If I say I am going to watch the match on Saturday
and then I do not, there seems to be something odd about calling
what I said false, even though I did not do what I said I would.
We feel that what makes an expression of intention true or false
is something about me now, while making a prediction is saying
something about the future.

Anscombe insists that this objection is wrongheaded. She wants
instead to focus on the fact that an expression of intention – ‘I
am going to watch the match on Saturday’ – is a description of
something future; as such, she says, it is true or false depending
on whether the description turns out true. She locates the source

EXPRESSIONS OF INTENTION FOR THE FUTURE 185

of the objection in a restricted conception of the ‘ways of saying
what is not true’, one which takes the concept of false observational
judgement as paradigm:

If I do not do what I said I would, I am not supposed to have made a
mistake, or even necessarily to have lied; so it seems that the truth of a
statement of intention is not a matter of my doing what I said. By why
should we not say: this only shows that there are other ways of saying
what is not true, beside lying and being mistaken?

(§2, p. 4)

Anscombe’s suggestion is that our ‘natural objection’ to saying
that an expression of intention for the future is false if I do not
do what I said I would stems from a prejudice about the ‘ways of
saying what is not true’. When I make an estimate of the future, if
what I say is false then this implies either that I was lying or that I
made a mistake. So, Anscombe’s suggestion is that there is a kind
of ‘saying what is not true’ that goes with expressions of intention
that does not fall into either of these categories.

Anscombe, anticipating the discussion of practical knowledge says:

[I]n some cases the facts are, so to speak, impugned for not being in
accordance with the words, rather than vice versa. This is sometimes
so when I change my mind; but another case of it occurs with e.g. I
write something other than what I think I am writing: as Theophrastus
says … the mistake here is one of performance, not of judgment. There
are other cases too: for example, St. Peter did not change his mind
about denying Christ, and yet it would not be correct to say he made a
lying promise of his faithfulness.

(§2, pp. 4–5)

So far as the ‘natural’ objection goes, any one of the three cases
Anscombe identifies here can illustrate her point: that there are
ways of saying what is not true which do not involve an error in
what was said. This point is what is required to undercut the source
of the intuition that expressions of intention cannot be a species
of prediction, as Anscombe wants to insist that they are. One of
the cases Anscombe describes – the one involving writing – is, of

EXPRESSIONS OF INTENTION FOR THE FUTURE186

course, by now familiar. It is the case in which I have practical
knowledge I am writing but, due to an error in action, what I am
doing is not happening.

This explains, then, why Anscombe breaks off at §3 to return to
her topic in §50. §3 answered the question – ‘[W]hat is the differ-
ence between e.g. “I am going to be sick” as it would most usually
be said, and “I am going to take a walk”, as it would most usually
be said?’ (§2, p. 1). The answer was: they are species of prediction,
but the former is justified by a reason for acting and the latter by
evidence. However, this description was ‘mystifying’ without an
understanding of ‘reason for acting’ and of the sort of error in act-
ing that would engage with the ‘deep’ objection we have to calling
‘I am going to do such-and-such’ a prediction. Having completed
these tasks Anscombe says: ‘what I have said about intention in
acting applies to intention in a proposed action’ (§50, p. 90).

Here is a good place to note how untroubled Anscombe need
be by the idea of a ‘state of pure intending’. The criteria which
determine whether a prediction is an expression of intention,
an order, or an estimate, have nothing to do with psychological
states, nor with whether a person has gone through a calculating
process. To say that an expression is an expression of intention is to
say that it could be justified by a reason for acting, and not by a
reason for thinking true.

‘I AM GOING TO DO IT UNLESS…’

Note that in that passage just quoted, Anscombe identifies two
other ways in which ‘the facts are … impugned for not being in
accordance with the words’. These apply to expressions of inten-
tion for the future but have no analogue in intention with which.
The first is the case in which I change my mind; the second is
when I do not do what I said I would.

Anscombe’s worry in §52 seems to be that these cases threaten
once again the idea that expressions of intention for the future are
a species of prediction. When I say ‘I am going to do such-and-
such’, it is always possible that I will change my mind or that I
will, in some way, be prevented from doing what I intend to do.
Thus it can start to look again as if I can’t be making a prediction

EXPRESSIONS OF INTENTION FOR THE FUTURE 187

when I express my intention, but must rather be saying some-
thing about myself now. And then we are back with the problem
of finding something at the time that an expression of intention
could be an expression of.

On Anscombe’s view, ‘I am going to do it but I will not do
it’ – where the first conjunct is an expression of intention and the
second a statement of belief – does not involve a ‘head-on’ contra-
diction (§52, p. 91). It is not like ‘I intend to do it and I intend
not to do it’. Nevertheless, it does involve a contradiction of sorts.
The contradiction, of course, is that ‘if the man does not go for a
walk, the first prediction is verified and the second falsified, and
vice versa if he does not go’ so there is real difficulty in under-
standing what such a man might mean (§52, p. 92). Prima facie,
then, a man makes himself unintelligible if I say ‘I intend to go
for a walk but I shall not go for a walk’. This is precisely the kind
of case we discussed in Chapter 6: ‘I am replenishing the water
supply but I am not replenishing the water supply’ is not a head
on contradiction, but the conjuncts have opposite truth values.
However, Anscombe thinks that with expressions of intention for
the future – in contrast with expressions of intentional action – it
is sometimes possible to sustain this thought, ‘p and ~p’ without
becoming unintelligible.

To make this point, Anscombe marks a distinction between ‘I
am going to do it unless I am prevented or change my mind ’ and
‘I am going to do it unless I do not do it’. She argues, somewhat
paradoxically, that the former is ‘absurd’ and the latter sometimes
‘the right thing to say’, and this despite the fact that the former is
not equivalent to the contradiction we just considered while the
latter is.

Concerning the first – ‘I am going to do it unless I am pre-
vented or change my mind’ – Anscombe says that such a statement
would be absurd, ‘like putting “unless my memory deceives me”
after every report one gave of what had happened’ (§52, p. 92).
Expressions of intention for the future, like all predictions, can be
wrong. It can turn out that what I said would happen does not.
This acknowledgement is not a species of scepticism – it does not
imply that ‘one could be wrong in every case’ (§52, pp. 93) – but
simple recognition that, unlike God, we are fallible beings.

EXPRESSIONS OF INTENTION FOR THE FUTURE188

However, Anscombe thinks that there are cases in which it can
be right to say ‘I will do it unless I do not’ (§52, p. 93). Such a
statement is not a general point about fallibility, but is apt when
one ‘is considering the fact that one may not do what one is deter-
mined to do’ (§52, p. 93). ‘I am going to do it’ communicates
one’s intention; ‘unless I do not’, one’s prediction, founded on
evidence, that one may fail.

A man hanging by his fingers from a precipice may be as certain as
possible that he must let go and fall, and yet determined not to let
go … [A] man could be as certain as possible that he will break down
under torture, and yet determined not to break down. And St. Peter
might perhaps have calculated ‘Since he says it, it is true’; and yet said
‘I will not do it’.

(§52, p. 94)

In each of these cases it is the fact that ‘verification of predic-
tions awaits the event’ combined with the ‘determination’ to do
the thing in question, that makes it intelligible for the person
to continue to intend to do what he believes he will not. The
man predicts, on strong grounds, that he will break under tor-
ture. But at the same time knows that the prediction may be
wrong – perhaps he will surprise himself, be braver, more resil-
ient, more determined than he supposed. So, given the possibility
of affecting the future, and the strength of his desire not to break
down, his intention remains. He stays committed to an idea of
how things should be and an idea of himself as an agent able to
make that the case. Of course, the more certain he is that he will
break down – if, for example, he has been tortured before, and
has broken – then the more ‘I will not break down’ becomes a
hope (§50, p. 50).

Of the St Peter case, Anscombe says: ‘The possibility in this
case arises from ignorance as to the way in which the prophesy
would be fulfilled’ (§52, p. 94). This ignorance is important.
Contrast a man who intends to replenish the water supply but
believes that he will not because there is a hole in the pipe. Here,
he knows precisely how the prediction ‘No you won’t’ is to be ful-
filled, and this must alter his knowledge and opinion about what

EXPRESSIONS OF INTENTION FOR THE FUTURE 189

can be the case and what can happen. If he does not thereby adopt
a different course, he becomes unintelligible. In contrast, though
St Peter may be certain, on the evidence of the prophesy, that he
will not keep his promise, he is nonetheless ‘ignorant of the way in
which the prophesy will be fulfilled’. His practical reason tells him
that he can keep his promise, and without knowledge that would
contradict what he knows about what is the case, and what can
happen, his intention can remain.4

INTENTION AND THE FUTURE

Anscombe’s investigation begins and ends with St Peter’s ‘promise
of faithfulness’ to Jesus (§2, p. 5 and §52, p. 94). The promise,
‘I will never disown you’, is an expression of intention for the
future.5 This connection with promising brings us to the question
of the point of expressions of intention for the future, and with it
to the point of the concept of intention.

When Anscombe says that her §3 account of expressions of
intention is mystifying, one thing she highlights is that it ‘assigns no
role to these predictions’ (§3, p. 6). ‘People do in fact give accounts
of future events in which they are some sort of agents … and these
accounts are very often correct’; but the role of these accounts is
‘quite obscure’ (§3, p. 7). We can see what is missing when we look
at Anscombe’s description of orders:

An imperative will be a description of some future action, addressed to
the prospective agent, and cast in a form whose point in language is to
make the person do what is described. I say its point in the language,
rather than the purpose of the speaker, partly because the speaker
might of course give an order with some purpose quite other than that
it should be executed (e.g. so that it should not be executed), without
detriment to its being an order.

(§2, p. 3)

In this passage Anscombe distinguishes the point in language and
the purpose of the speaker; she is interested in describing the lin-
guistic practice of giving and obeying orders, not the psychology
of people engaged in that practice. Think of Little Big Man and

EXPRESSIONS OF INTENTION FOR THE FUTURE190

General Custer. General Custer knows that Little Big Man wishes
him dead, so uses him as a ‘reverse barometer’: ‘Anything that
man tells me will be a lie’. Little Big Man knows this, so when the
General asks him, ‘What should I do?’, Little Big Man reflects:
‘I had him. But this time what I held in my hand wasn’t a knife,
but the truth’, and answers with the imperative: ‘You go down
there!’. Custer, wise to Little Big Man’s knowledge responds: ‘You
want me to think that you don’t want me to go down there, …
but the subtle truth is you really don’t want me to go down there.’
Of course, Little Big Man wanted Custer to think that he (Little
Big Man) wanted him (Custer) to think that he (Little Big Man)
didn’t want him to go down there so that Custer would follow his
imperative. Custer duly does and meets a sticky end.

None of this undermines the fact that the point in language of
the imperative is to ‘make the person do what is described’, but
here the purposes of the speaker in issuing the imperative is far
from straightforward.

The difficulty at that point – §3 – was to see what the ‘point
in language’ could be of an expression of intention for the future.
I want to connect expressions of intention with two important
aspects of what Anscombe elsewhere identifies as necessary for
‘living in a society’ (Anscombe, 1969a, p. 18).

First, as we have seen, expressions of intention for the future are
intimately connected with the conventions of giving orders and
making promises, conventions which are ways of ‘getting human
beings to do things’ – or not to do them – which Anscombe else-
where identifies as an Aristotelian necessity:

What ways are there of getting human beings to do things? You can
make a man fall over by pushing him; you cannot usefully make his
hand write a letter or mix concrete by pushing; for in general if you
have to push his hand in the right way, you might as well not use him
at all. You can order him to do what you want, and if you have authority
he will perhaps obey you. Again if you have the power to hurt him or
help him according as he disregards or obeys your orders, or if he
loves you so as to accord with you requests, you have a way of getting
him to do things. However, few people have authority over everyone
they need to get to do things, and few people either have power to hurt

EXPRESSIONS OF INTENTION FOR THE FUTURE 191

or help others without damage to themselves or command affection
from others to such an extent as to be able to get them to do the
things they need others to do … [I]n default of the possibility or utility
of exerting physical force, and of the possibility of exercising authority
or power to hurt and help, or of commanding affection, this feeble
means is at least a means of getting people to do things. Now getting
one another to do things without the application of physical force is
a necessity for human life, and that far beyond what could be secured
for those other means.

(Anscombe, 1969a, p. 18)

Of course, not all expressions of intention for the future are prom-
ises; rather, an expression of intention is making a promise in
certain circumstances. Which is to say an answer to the question
‘Why did you say “I will do it”?’ can be ‘I was making a promise’.

Even more fundamental, we saw in §§20–21 that Anscombe
argues that without expressions of intention for the future, the
concept of intention would ‘lose its point’:

It seems reasonable to say that if the only occurrence of intention were
as the intention of doing whatever one is doing, the notion of inten-
tion would be a very thin one; it is not clear why it should be marked
off as a special class among all those of a man’s actions and move-
ments which are known to him without observation, any more than we
mark off movements that are expressions of emotion as a distinct and
important class of happenings.

(§20, p. 32)

So, the question we end with is: why would the concept of ‘inten-
tion’ lose its point in this way?

In two places in Intention (§14, p. 22; §20, pp. 31–32) Anscombe
draws attention to the practice of criticising a man for his motives.
For example, we say ‘That he killed your father didn’t give you
a reason to kill him’, or ‘You shouldn’t have let spiteful feelings
influence you in that way’. In both places Anscombe suggests that
this practice is most important when it is connected with proposed
actions rather than things a person has already done.

EXPRESSIONS OF INTENTION FOR THE FUTURE192

One can argue against motives – i.e. criticise a man for having acted
on such a motive – but a great deal of the point of doing so will be
gone if we imagine the expression of intention for the future absent.

(§20, p. 31)

This connects once again with our Aristotelian necessity. As I sug-
gested in Chapter 5: without the prospect of criticising a proposed
action, the practice of criticising a current one would lose its point.
When a person says what they will do, their proposed action is
introduced as an object for criticism and debate. I may decide I
want to do it too; I may decide to try and stop them, help them,
advise them, warn them, watch them. The possibility of doing
something together, of creating some shared reality opens up.

Anscombe, I think, goes further. She says that without expres-
sions of intention for the future:

It is not clear why [intentional action] should be marked off as a spe-
cial class among all those of a man’s actions and movements that
are known to him without observation, any more than we mark off
movements that are expressions of emotion as an important class of
happenings.

(§20, p. 32)

According to Anscombe, then, though the concept of intention is
most clearly elucidated by looking at intentional actions, the fine
distinctions that we have described are ones that would cease to
interest us were it not for the fact that we speak of intention in
proposed action. In this sense, then, ‘intention always concerns the
future’ (§1, p. 1).

NOTES

1 See Hursthouse (2000) for a discussion.
2 For an attempt to do just this see McDowell (1991). For arguments that

such an account necessarily fails, see Wright (1991).
3 I would suggest that this parallels the argumentative strategy employed

by Wittgenstein in Philosophical Investigations. It is a mistake to respond
to the ‘private language argument’ by trying to show that such a language

EXPRESSIONS OF INTENTION FOR THE FUTURE 193

is possible; the right place to engage is with the account of the grammar
of our sensation words.

4 For an illuminating and detailed discussion of the relation between
promising and believing see Berislav Marušić (2015). If Anscombe is
right, then promising that I will A does not require that I believe I will
A. At least, not in those cases in which one is determined to A, but has
reason to believe A-ing will not be possible.

5 From the Matthew, 26:31–35:

Then Jesus told them, ‘This very night you will all fall away on
account of me, for it is written:

“I will strike the shepherd, and the sheep of the flock will be
scattered.”

But after I have risen, I will go ahead of you into Galilee.’
Peter replied, ‘Even if all fall away on account of you, I never will.’
‘Truly I tell you,’ Jesus answered, ‘this very night, before the

rooster crows, you will disown me three times.’ But Peter declared,
‘Even if I have to die with you, I will never disown you.’ And all the
other disciples said the same.

SUGGESTED READING

Anscombe, G. E. M. (1969a). ‘On Promising and its Justice, and Whether it Need
be Respected in Foro Interno’. Critica (Revisita Hispanoamerica de FilosofiaI)
(Mexico), 3, 7/8. Reprinted in ERP, pp. 10–21.

Foot, Philippa (2001). Natural Goodness (Oxford: Oxford University Press), chapter 3.
Hamilton, Andy (2008). ‘Intention and the Authority of Avowals’, Philosophical

Explorations, 11(1), pp. 23–37.
Vogler, Candace (2002) Reasonably Vicious (Cambridge, Mass.: Harvard University

Press).
—— (2012) ‘In Support of Moral Absolutes’, Villanova Law Review, 57(5), pp. 893–906.
Wright, Crispin (1991). ‘Wittgenstein’s Later Philosophy of Mind: Sensation, Privacy

and Intention’. In Klaus Puhl (ed.), Meaning Scepticism (Berlin and New York:
Walter de Gruyter), pp. 126–147.

WORKS BY ELIZABETH ANSCOMBE

1939. ‘The Justice of the Present War Examined’. Pamphlet published by authors.
Reprinted in ERP, pp. 72–81.

1950. ‘The Reality of the Past’. Reprinted in MPM, pp. 103–119.
1957a. Intention (Oxford: Basil Blackwell). 2nd edition, 1972.
1957b. Does Oxford Moral Philosophy Corrupt the Youth?’. Radio broadcast,

printed in The Listener, Vol. 57 (14 February), pp. 226–227. Reprinted in HLAE,
pp. 161–168.

1957c. ‘Mr Truman’s Degree’. Pamphlet published by author. Reprinted in ERP, pp. 62–71.
1957d. ‘Intention’. Proceedings of the Aristotelian Society. Reprinted in MPM, pp. 75–82.
1957d. ‘Oxford Moral Philosophy’. In The Listener, no. 1457, 28 February (London,

England).
1958a. ‘Pretending’. Proceedings of the Aristotelian Society, supplementary volume 32.

Reprinted in MPM, pp. 83–93.
1958b. ‘Modern Moral Philosophy’. Philosophy, 53, pp. 1–19. Reprinted in HLAE,

pp. 169–194.
1958c. ‘On Brute Facts’, Analysis, 18(3). Reprinted in ERP, pp. 22–25.
1961. ‘War and Murder’. Walter Stein (ed.), Nuclear Weapons: A Catholic Response

(London and New York). Reprinted in ERP, pp. 51–61.
1962a. ‘Authority in Morals’. John Todd (ed.), Problems of Authority (London: Darton,

Longman and Todd). Reprinted in ERP, pp. 43–50.
1962b. ‘On Sensations of Position’, Analysis, 22( 3). Reprinted in MPM, pp. 71–74.
1963a. ‘The Two Kinds of Error in Action’, Journal of Philosophy, 60. Reprinted in ERP,

pp. 3–9.

BiBliogrAphy

BIBLIOGRAPHY 195

1963b. ‘Events in the Mind’. Printed in MPM, pp. 57–63.
1964. ‘Substance’. Proceedings of the Aristotelian Society: Supplementary volume 38.

Reprinted in MPM, pp. 37–43.
1965a. ‘Thought and Action in Aristotle: What is “Practical Truth”?’. From J. R. Banbrough

(ed.), New Essays on Plato and Aristotle (London). Reprinted in FPW, pp. 66–77.
1965b. ‘The Intentionality of Sensation: A Grammatical Feature’. From R. J. Butler

(ed.), Analytical Philosophy, second series (Oxford). Reprinted in MPM, pp. 3–20.
1968. ‘You Can have Sex without Children: Christianity and the New Offer’. Renewal

of Religious Structures: Proceedings of the Canadian Centenary Theological Congress
(Toronto). Reprinted in ERP, pp. 82–96.

1969a. ‘On Promising and its Justice, and Whether it Need be Respected in Foro
Interno’. Critica (Revisita Hispanoamerica de FilosofiaI) (Mexico), 3, 7/8. Reprinted
in ERP, pp. 10–21.

1969b. ‘Causality and Extensionality’. Reprinted in MPM, pp. 173–179.
1971. ‘Causality and Determination’. Reprinted in MPM, pp. 133–147.
1974a. ‘Practical Inference’. From P. A. Schilpp and L. E. Hahn (eds), The Philosophy

of Georg Henrik von Wright. Library of Living Philosophers Series Volume XIX
(La Salle, Ill: Open Court 1989). Reprinted in HLAE, pp. 109–148.

1974b. ‘The First Person’. From Samuel Guttenplan (ed.), Mind and Language:
Wolfson College Lectures 1974 (Oxford). Reprinted in MPM, pp. 21–36.

1974c. ‘Times, Beginnings, and Causes’. Reprinted in MPM, pp. 148–162.
1974d. ‘Memory, “Experience”, and Causation’. Reprinted in MPM, pp. 120–132.
1974e. ‘Comments on Professor R. L. Gregory’s Paper on Perception, Perception’,

in S. C. Brown (ed.), Philosophy of Psychology (London: Macmillan). Reprinted in
MPM.

1975. ‘The First Person’. Reprinted in MPM, pp. 21–36.
1976a. ‘The Question of Linguistic Idealism’. From ‘Essays on Wittgenstein in hon-

our of G. H. von Wright’, Acta Philosophica Fennica, 28, 1–3. Reprinted in FPW,
pp. 112–133.

1976b. ‘The Subjectivity of Sensation’. Ajatus (Yearbook of the Philosophical Society
of Finland), 36. Reprinted in MPM, pp. 44–56.

1977. ‘Soft Determinism’. Reprinted in MPM, pp. 163–172.
1978. ‘On the Source of the Authority of the State’. Reprinted in ERP, pp. 130–155.
1979a. ‘Chisholm on Action’. From Grazer Philosophisch Studien (Ernest Sosa (ed.),

Essays on the Philosophy of Roderick M. Chisholm) 7/8, pp. 206–213. Reprinted in
HLAE, pp. 77–89.

1979b. ‘“Under a Description”’. Nous, 13. Reprinted in MPM, pp. 208–219.
1981. ‘Commentary on John Harris’ “Ethical Problems in the Management of Severely

Handicapped Children”’. From Journal of Medial Ethics, 7, pp. 122–123. Reprinted
in HLAE, pp. 279–284.

1982a. ‘Making True’. Roger Teichmann (ed.), Logic, Cause, and Action. Cambridge
University Press (2000), pp. 1–8.

1982b. ‘Action, Intention, and “Double Effect”’. Proceedings of the American Catholic
Philosophical Association, Vol. 56, pp. 12–25. Reprinted in HLAE, pp. 207–226.

1982c. ‘Murder and the Morality of Euthanasia’. Printed in HLAE, pp. 261–277.

BIBLIOGRAPHY196

1983a. ‘The Causation of Action’. From C. Ginet (ed.), Knowledge and Mind (New York
and Oxford: Oxford University Press), pp. 174–190. Reprinted in HLAE, pp. 89–108.

1983b. ‘Sins of Omission? The Non-Treatment of Controls in Clinical Trials’.
Proceedings of the Aristotelian Society: Supplementary Volume, 57, pp. 223–227.
Reprinted in HLAE, pp. 285–291.

1989. ‘Sin’. Reprinted in FHG, pp. 117–156.
1993. ‘Practical Truth’. From John M. Dolan (ed.), Working Papers in Law, Medicine,

and Philosophy (Program in Human Rights and Medicine of the University of
Minnesota). Reprinted in HLAE, pp. 149–158.

Undated a. ‘Good and Bad Human Action’. Undated Manuscript. Printed in HLAE,
pp. 195–206.

Undated b. ‘The Moral Environment of the Child’. Printed in FHG, pp. 224–232.

COLLECTIONS BY ANSCOMBE

PW. From Parmenides to Wittgenstein: Collected Philosophical Papers Volume I (Blackwell:
1981).

MPM. Metaphysics and the Philosophy of Mind: Collected Philosophical Papers Volume II
(Blackwell: 1981).

ERP. Ethics, Religion and Politics. Collected Philosophical Papers Volume II (Blackwell:
1981).

HLAE. Mary Geach and Luke Gormally (eds), Human Life, Action and Ethics (Imprint
Academia: 2005).

FHG. Mary Geach and Luke Gormally (eds), Faith in a Hard Ground: Essays on
Religion, Philosophy, and Ethics by G.. E. M. Anscombe (Imprint Academic: 2008).

PlW. Mary Geach and Luke Gormally (eds), From Plato to Wittgenstein: Essays by
G.. E. M. Anscombe (Imprint Academia: 2005).

OTHER WORKS

Austin, J. L. (1957). ‘A Plea for Excuses’, Proceedings of the Aristotelian Society,
Vol. 57.

—— (1958). ‘Pretending’, Proceedings of the Aristotelian Society, Supplementary
Volumes, Vol. 32, pp. 261–278.

Baker, Gordon (2001). ‘Wittgenstein’s “Depth Grammar”’. Reprinted in Katherine J.
Morris (ed.), Wittgenstein’s Method: Neglected Aspects. Essays on Wittgenstein by
Gordon Baker (Oxford: Blackwell, 2004), pp. 73–91.

Baier, Kurt (1960) ‘Critical Notice’, Australisian Journal of Philosophy, 38(1), pp. 71–81.
Blackburn, Simon (2005a). ‘Simply Wrong’. Times Literary Supplement, Friday 30

September, Issue 5347.
—— (2005b). ‘Anscombe’s Ethics’ in letters, The Times Literary Supplement (London,

England), Friday 14 October, Issue 5349, p. 19.
—— (2005c). ‘Anscombe’s Ethics’ in letters, The Times Literary Supplement (London,

England), Friday 21 October, Issue 5350, p. 19.

BIBLIOGRAPHY 197

Bratman, Michael (1987). Intention, Plans, and Practical Reason (Cambridge, Mass.:
Harvard University Press).

Chisholm, R. M. (1959). ‘Review’. The Philosophical Review, 68(1; January).
Conradi, Peter J. (2001). Iris Murdoch: A Life. Harper Collins.
Cowen, Joseph (1959). Ethics, 69(2; January), pp. 144–151.
Dancy, Jonathan (2010). ‘Various debts’, Times Literary Supplement (April).
Davidson, Donald (1963). ‘Actions, Reasons, and Causes’. In his Essays on Actions

and Events. Oxford: Oxford University Press (2001), pp. 3–20.
—— (1970a). ‘How is Weakness of the Will Possible?’. In his Essays on Actions and

Events. Oxford University Press (2001), pp. 21–42.
—— (1970b). ‘Mental Events’. In his Essays on Actions and Events. Oxford University

Press (2001), pp. 207–224.
—— (1971). ‘Agency’. In his Essays on Actions and Events. Oxford University Press

(2001), pp. 43–62.
—— (1978). ‘Intending’. In his Essays on Actions and Events. Oxford University Press

(2001), pp. 83–102.
Denyer, Nicholas (1997). ‘Is Anything Absolutely Wrong?’. In David S. Oderberg

and Jacqueline A. Laing (eds), Human Lives: Critical Essays on Consequentialist
Bioethics (Palgrave Macmillan), pp. 39–57.

Diamond, Cora (1966). ‘Secondary Sense’, Proceedings of the Aristotelian Society 67.
Reprinted in The Realistic Spirit, pp. 225–242.

—— (1977). ‘Riddles and Anselm’s Riddle’, Aristotelian Society Supplementary Volume
51, pp.143–186.

—— (1991). The Realistic Spirit: Wittgenstein, Philosophy, and the Mind (Cambridge,
Mass.: MIT Press).

—— (1997). ‘Consequentialism in Modern Moral Philosophy and in “Modern
Moral Philosophy”’. In David S. Oderberg and Jacqueline A. Laing (eds),
Human Lives: Critical Essays on Consequentialist Bioethics (Palgrave Macmillan),
pp. 13–38.

Driver, Julia (2011). ‘Gertrude Elizabeth Anscombe’. In Stanford Encyclopaedia of
Philosophy.

Falvey, K. (2000). ‘Knowledge in Intention’, Philosophical Studies, 99, pp. 21–44.
Finkelstein, David (2009). Expression and the Inner (Cambridge, Mass.: Harvard

University Press).
Foot, Philippa (2001). Natural Goodness (Oxford: Oxford University Press).
Ford, Anton (2011). ‘Action and Generality’. In Anton Ford, Jennifer Hornsby and

Frederick Stoutland (eds), Essays on Anscombe’s Intention (Cambridge, Mass.:
Harvard University Press), pp. 76–104.

Frege, Gottlob (1879). Begriffsschrift.
—— (1956). ‘The Thought: A Logical Enquiry’, Mind, 259(July), pp. 289–311.
—— (1984). Collected Papers on Mathematics, Logic, and Philosophy. Edited by Brian

McGuinness (Oxford: Oxford University Press).
Geach, Mary (2005). ‘Introduction’. Human Life, Action and Ethics (Imprint Academia).
—— (2011). ‘Introduction’. From Plato to Wittgenstein: Essays by G. E. M. Anscombe.

Edited by L. Geach and M. Gormally (Imprint Academia).

BIBLIOGRAPHY198

Geach, Peter (1958). Mental Acts: Their Content and Their Objects (London: Routledge
and Paul).

—— (2000). ‘Intention, Freedom, and Predictability’, in Roger Teichmann (ed.),
Logic, Cause and Action (Cambridge: Cambridge University Press), pp. 73–82.

Gellner, Ernest (1959). Words and Things: A Critical Account of Linguistic Philosophy
and a Study in Ideology (London: Gollancz; Boston: Beacon).

Goldman, Alvin (1970). A Theory of Human Action (Englewood Cliffs, NJ: Prentice-
Hall).

Gustafson, Don F. (1964). ‘Voluntary and Involuntary’, Philosophy and Phenomeno­
logical Research, 24(4), pp. 493–501.

Haddock, Adrian (2011). ‘“The Knowledge That a Man has of His Intentional
Actions”’. In Anton Ford, Jennifer Hornsby and Frederick Stoutland (eds), Essays on
Anscombe’s Intention (Cambridge, Mass.: Harvard University Press), pp. 147–169.

Haldane, John (2004). ‘Thomism and the future of Catholic philosophy’, in his
Faithful Reason: Essays Catholic and Philosophical (London: Routledge).

Hamilton, Andy (2008). ‘Intention and the Authority of Avowals’, Philosophical
Explorations, 11(1), pp. 23–37.

—— (2013). The Self in Question: Memory, the Body, and Self­Consciousness (London:
Palgrave).

Hare, R. M. and Nowell-Smith, P. H. (1957). ‘Oxford Moral Philosophy’, in The
Listener, no. 1456, 21 February, p. 311 (London, England).

Heath, P. L. (1960). ‘Review’, in The Philosophical Quarterly, 10(40: July).
Hornsby, J. (2011). ‘Actions and Their Circumstances’. In Anton Ford, Jennifer

Hornsby and Frederick Stoutland (eds), Essays on Anscombe’s Intention
(Cambridge, Mass.: Harvard University Press), pp. 105–127.

Hume, David (1739/2007). A Treatise of Human Nature: Volume 1. Edited by David
Fate Norton and Mary J. Norton (Clarendon Hume Edition Series, 2007).

Hursthouse, Rosalind (1999). On Virtue Ethics (Oxford: Oxford University Press).
—— (2000). ‘Intention’. In Roger Teichmann (ed.), Logic, Cause, and Action

(Cambridge: Cambridge University Press), pp. 83–106.
Jarvis, Judith (1959). ‘Review’, The Journal of Philosophy, 56(1; 1 January).
Kennedy, David M. (1999). The American People in World War II (Oxford: Oxford

University Press).
McDowell, John (1991). ‘Intentionality and Interiority in Wittgenstein’, reprinted

in his Mind, Value, and Reality (Cambridge, Mass.: Harvard University Press
(1998)), pp. 297–321.

—— (1994). Mind and World (Cambridge, Mass.: Harvard University Press).
—— (1998). Mind, Value, and Reality (Cambridge, Mass.: Harvard University Press).
—— (2011). ‘Anscombe on Bodily Self-Knowledge’. In Anton Ford, Jennifer Hornsby

and Frederick Stoutland (eds), Essays on Anscombe’s Intention (Cambridge,
Mass.: Harvard University Press), pp. 128–146.

McGinn, Marie (1997). Routledge Philosophy Guidebook to Wittgenstein and the
Philosophical Investigations (London: Routledge).

—— (2013). Routledge Philosophy Guidebook to Wittgenstein and the Philosophical
Investigations, 2nd edition (London: Routledge).

BIBLIOGRAPHY 199

MacIntyre, Alisdair (1990). Three Rival Versions of Moral Enquiry: Encyclopaedia,
Genealogy, and Tradition (London: Duckworth).

Martin, C. B. (1971). ‘Knowledge without Observation’, Canadian Journal of Philosophy,
1(1; September), pp. 15–24.

Marušić, Berislav (2015). Evidence and Agency: Norms of Belief for Promising and
Resolving (Oxford: Oxford University Press).

Mehta, Ved (1965). Fly and the Fly­Bottle (London: Pelican Books).
Midgley, Mary. (2005).The Owl of Minerva: A Memoir (London: Routledge).
—— (2013). Letter published in Guardian (28 November), available at: http://

www.theguardian.com/world/2013/nov/28/golden-age-female-philosophy-
mary-midgley (accessed on 8 September 2015).

Moore, G. E. (1903). Principia Ethica (Cambridge: Cambridge University Press).
Moran, Richard (2004). ‘Anscombe on “Practical Knowledge”’. In J. Hyman and

H. Steward (eds), Royal Institute of Philosophy Supplement (Cambridge: Cambridge
University Press), pp. 43–68.

Moran, Richard and Stone, Martin J. (2011). ‘Anscombe on Expressions of Intention:
An Exegesis’. In Anton Ford, Jennifer Hornsby and Frederick Stoutland (eds), Essays
on Anscombe’s Intention (Cambridge, Mass.: Harvard University Press), pp. 33–75.

Mounce, Howard (2011). ‘Winch and Anscombe on Ethics and Religion’. Philosophical
Investigations, 34(3), pp. 241–248.

Müller, Anselm Winfred (2011). ‘Backward-Looking Rationality and the Unity of Practical
Reason’. In Anton Ford, Jennifer Hornsby, and Frederick Stoutland (eds), Essays on
Anscombe’s Intention (Cambridge, Mass.: Harvard University Press), pp. 242–269.

Nichols, John (1808). The Works of the Rev. Jonathan Swift, D.D. Dean of St. Patrick’s,
Dublin. Arranged by Thomas Sheridan, A.M. With notes, Historical and Critical, Vol. I
(London: J. Johnson, J. Nichols and Son).

O’Brien, Lucy F. (1994). ‘Anscombe and the Self-Reference Rule’. Analysis, 54(4),
pp. 277–281.

Pickard, Hanna (2004). ‘Knowledge of Action without Observation’, Proceedings of
the Aristotelian Society, 104, pp. 203–228.

Reid, Thomas (1788/2010). Essays on the Active Powers of Man. Edited by Knud Haakonssen
and James A. Harris (University Park: Pennsylvania State University Press).

Richter, Duncan (2011). Anscombe’s Moral Philosophy (Plymouth: Lexington).
Rorty, Richard (1967). ‘Metaphilosophical Difficulties of Linguistic Philosophy’, in

The Linguistic Turn: Recent Essays in Philosophical Method (Chicago and London:
University of Chicago Press), pp. 1–39.

Russell, Bertrand (1905). ‘On Denoting’, Mind, 14, pp. 479–493.
Ryle, Gilbert (1949). The Concept of Mind (London: Hutchinson).
Schwenkler, John (2015). ‘Understanding “Practical Knowledge”’, in Philosophers’

Imprint, 25(15; June).
Searle, John (1983). Intentionality: An Essay in the Philosophy of Mind (Cambridge:

Cambridge University Press).
Setiya, Kieran (2011). ‘Knowledge of Intention’. In Anton Ford, Jennifer Hornsby and

Frederick Stoutland (eds), Essays on Anscombe’s Intention (Cambridge, Mass.:
Harvard University Press), pp. 170–197.

http://www.theguardian.com/world/2013/nov/28/golden-age-female-philosophy-mary-midgley
http://www.theguardian.com/world/2013/nov/28/golden-age-female-philosophy-mary-midgley
http://www.theguardian.com/world/2013/nov/28/golden-age-female-philosophy-mary-midgley

BIBLIOGRAPHY200

—— (2014) ‘Intention’. Stanford Encyclopedia of Philosophy.
Stoutland, Frederick (2011). ‘Anscombe’s Intention in Context’. Anton Ford,

Jennifer Hornsby and Frederick Stoutland (eds), Essays on Anscombe’s Intention
(Cambridge, Mass.: Harvard University Press).

Strawson, P. F. (1959). Individuals: An Essay in Descriptive Metaphysics (London:
Methuen).

—— (1960). ‘The Post-Linguistic Thaw’. The Times Literary Supplement (London,
England), Friday, 9 September, p. 584, Issue 3054.

—— (1966). The Bounds of Sense (London: Methuen).
Teichman, Jenny (2002). ‘Gertrude Elizabeth Margaret Anscombe: 1919–2001’,

Biographical Memoirs of Fellows, Vol. 115 (Oxford: Oxford University Press),
pp. 30–50.

Teichmann, Roger (2008). The Philosophy of Elizabeth Anscombe (Oxford: Oxford
University Press).

—— (2014a) ‘Why “Why?” Action, Reasons and Language’, in Philosophical Investiga­
tions, pp. 1–18.

—— (2014b). ‘The Voluntary and the Involuntary: Themes from Anscombe’,
American Catholic Philosophical Quarterly, 88(3), pp. 465–486.

Thompson, Michael (2008). Life and Action: Elementary Structures of Practice and
Thought (Cambridge, Mass.: Harvard University Press).

—— (2011). ‘Anscombe’s Intention and Practical Knowledge’. In Anton Ford,
Jennifer Hornsby and Frederick Stoutland (eds), Essays on Anscombe’s Intention
(Cambridge, Mass.: Harvard University Press), pp. 198–210.

Trethovvan, Illtyd (1958). ‘The Ethics of Intention’, The Tablet, 11 January, p. 12.
Vesey, G. N. A. (1963). ‘Knowledge without Observation’, The Philosophical Review,

72, pp. 198–212.
Vogler, Candace (2002). Reasonably Vicious (Cambridge, Mass.: Harvard University

Press).
—— (2012). ‘In Support of Moral Absolutes’, Villanova Law Review, 57(5), pp. 893–906.
Von Wright, G. H. (1972). ‘On So-called Practical Inference’, Acta Sociologica, 15/1.
Winch, Peter (1976). ‘Im Anfang war die Tat’, in his Trying to Make Sense (Oxford:

Blackwell, 1987), pp. 33–53.
—— (1986). ‘Who is my Neighbour?’, in his Trying to Make Sense (Oxford: Blackwell,

1987), pp. 154–166.
Wittgenstein, L. (1922). Tractatus Logico­Philosophicus (London: Kegan Paul).
—— (1958). Philosophical Investigations. G. E. M. Anscombe (trans.) Second edition

(Oxford: Basil Blackwell).
—— (1967). Zettel. G. E. M. Anscombe and Georg Henrik von Wright (eds),

University of California Press.
—— (1980). Culture and Value (Chicago, Ill.: University of Chicago Press).
Wright, Crispin (1991). ‘Wittgenstein’s Later Philosophy of Mind: Sensation, Privacy

and Intention’. In Klaus Puhl (ed.), Meaning Scepticism (Berlin and New York:
Walter de Gruyter), pp. 126–147.

absolutism see ethics of prohibition
action: bodily 21–2; human 27, 40–45,

83, 93, 152–3, 157–8; intentional
7, 21–2, 44–5, 48, 50, 73, 77–112,
114–20, 124–8; involuntary 15, 73,
79, 81–3, 84, 93, 101; voluntary 79,
83, 106–9

Action Identity Thesis 3n2, 21–2, 23
analytic: Aristotelianism 23; philosophy

23; Thomism 23
animal 61, 156–7
Anscombe’s Question see connective

approach
Aristotle 18, 23, 27, 35–6, 65, 119, 132–3,

146–7, 174
Aristotelian necessity 179, 190–2
Austin, J. L. 13, 14–17, 63, 71

Baker, Gordon 16n2, 64n6
behaviourism 54, 160
belief 187–9
Blackburn, Simon 35–6
bodily awareness see knowledge of

bodily posture
Bratman, Michael 52

calculative order 113–14, 116, 123–8,
152–3, 156, 161

capacity to say 91–2
Catholicism: Anscombe’s 7, 10, 12,

17–20; and ethics 33–6, 39, 82, 107,
128, 140; and form of life 2, 147; see
also doctrine of double effect

Cartesian psychology 40, 47, 52, 85,
128–32, 158–61, 172–3

causation: mental 3, 22–3, 79, 80,
92–7, 98, 104, 119; and responsibility
30–31

cause vs reason 22–3, 79, 81–3, 92–7,
98, 99–103, 132, 183; see also
causation and reason for acting

clarity 60, 69–70, 77, 78–9, 83, 106, 134,
151, 154

cliché 2
command 163–9, 183, 189
connective approach 49–56, 63, 114,

152, 160, 178
consequences: intended vs foreseen

27, 37–9, 82–3; relevance for moral
evaluation 29, 31, 32–6, 37–9, 82–3;
see also consequentialism

index

INDEX202

consequentialism 32–6
corruption 17, 32

Dancy, Jonathan 1
Davidson-Anscombe Thesis see Action

Identity Thesis
Davidson, Donald 3, 22, 51–3, 67, 82,

108, 160
description under which see under a

description
desire: feeling of 31, 104–6, 146; and

intention 31, 104–6, 145–7; object of
104–6, 143–4, 146–7

deontological ethics see ethics
Diamond, Cora 2, 3n4, 4n5, 5n6, 33n3,

87n5, 109, 114n1, 124
doctrine of double effect 39–40, 47,

128–32

error 90–1, 163–74, 179, 185–6
ethics: deontological 34–6, 140; of

prohibition 33–4, 37, 128; virtue
27, 36, 40–45, 47; see also moral
philosophy

evil 29
expression of intention 7, 38, 73–4,

115–120, 153, 155, 163, 166–7, 169,
177–93

fact: brute 43–4, 127; vs value see Hume
falsehood see truth
Frege, Gottlob 3, 4, 13–14, 119
friendship 103
Foot, Philippa 12, 26, 36n5
form of description of events 152–8,

160–1, 179
formal order 3, 4, 7, 73–4, 106, 120,

124–5, 132, 146, 152–3, 156, 180; see
also calculative order

Geach, Peter 10, 59n4
Gellner, Ernest 13n1, 16
Goldman, Alvin 3n2
good 80, 100–3, 147, 179, 183
grammar 3, 5, 16, 64, 74, 159, 160,

166–7, 180; see also method

Haddock, Adrian 135n5
Haldane, John 23
Hamilton, Andy 84n4
Hare, R. M. 13
harm 80, 100–3, 183
human: action 27, 40–45, 93, 152, 157–8;

identity 67–8; life 2, 4, 36, 42, 45,
58–62, 109, 119, 127, 146–7, 155, 179,
180, 190–2

Hume: and causation 23, 82, 94–6, 102;
and is/ought gap 41–45

Hursthouse, Rosalind 8, 81n1, 135n5,
181n1

identity: action 21, 67–9, 124–8; event
67–8; human 67–8; see also Action
Identity Thesis

intelligibility 108–9, 121–2, 130–1, 154–5,
165–6, 169, 187–9

intention with which 7, 27, 39–40, 50,
73, 80, 113–50

inner: act of mind 40, 130–1;
impression 146

intuition 57–8, 69–70, 77, 99, 114, 133,
138, 151, 159, 160, 162, 180, 183

knowledge: of bodily posture 3, 84–6,
88, 90–1, 134; and error 91–2, 173–4;
object of 89, 135–8; by observation
79, 84–92, 165–6, 169, 184; under
a description 79, 135–8; without
observation 74, 83–92, 93–4, 99,
118, 132–8, 162, 172–4; see also
practical knowledge and speculative
knowledge

language, conception of 3
linguistic philosophy 13–17, 23, 42,

175; see also ordinary language
philosophy

linguistic practice 58–9, 69, 78, 179,
180, 182

logical form 3, 14
logical compulsion 139–40, 141–5,

162–3
lying 131, 185

INDEX 203

Martin, C. B. 86
Marušić, Berislav 189n4
Mehta, Ved 16
mankind: natural history of 2,

12, 42, 109, 127, 147; see also
human life

mental: cause see causation; process 4,
113, 128, 156, 186; state 3, 50, 51–4,
181, 186

metaphilosophy see method
metaphysics 13, 14–15
McDowell, John 48, 72, 181n2
method 4–6, 7, 13–17, 56–65, 114, 180,

182, 184
Midgley, Mary 12
mind–body problem 85–8, 172–3
Moore, G. E. 42, 155n1
moral philosophy: Anscombe’s 19;

critique of modern 6, 26–7, 31–6, 42,
65; see also ethics

Moran, Richard 135n5
motive 81, 98, 99–105, 116–17, 183,

191–2
movement, involuntary 92–7, 118–19
murder 30–1, 130
Murdoch, Iris 11, 12

neo-Kantian 47–8
nuclear weapons see World War II

object: of desire 89, 104–6, 143–4,
146–7; grammatical 89; intentional
89–90; of sensation 89–90; of
knowledge 89, 135–8

observation see knowledge by
observation

order see command or formal
order

ordinary language philosophy 14–17, 71

pacifism 19, 30
Pickard, Hanna 87
primitive 100
practical knowledge 138, 151, 161–75, 185
practical reason 50, 74, 92, 105, 113,

132–3, 138, 152–3, 161–2, 170, 172;

vs. theoretical reason 138–45, 162,
174, 178

prediction 38, 99, 116, 153, 155, 178–86
pretence 131
progressive, the 7n7
promising 185, 188–92
psychological: account of

intention 2, 4, 109, 117–19, 121,
123–4, 128, 142–6, 159–61, 180–1, 184;
mechanism 3, 145–6; process 3, 186

psychology, philosophy of 6, 19, 27,
41, 128

pure intending 51–2, 186

reason for action 22–3, 78–9, 80, 81–3,
92–7, 98–103, 178, 184, 186; see also
cause vs reason

regret 83
responsibility 37–8; and causation

30–31
revenge 100–1
Russell, Bertrand 16, 89

Saint Peter 55, 185, 188–9
Saint Thomas Aquinas 18–19, 23, 27, 65,

83n2, 83n3, 88n6, 174–5
Schwenkler, John 18, 88n6
separately describable sensations 84,

86–92
Setiya, Kieran 22, 50
self-knowledge 85–8
sense-data 89
speculative knowledge 161, 171, 173–4
Strawson, P. F. 13, 16, 23, 47–8

Teichman, Jenny 17
Teichmann, Roger 5n6
Theophrastus 171–2, 179, 185
theoretical reasoning 138–45
Thompson, Michael 23, 36n5, 55, 91
Truman, Harry S. 26–31, 37, 129
Truth 138, 162, 168–9, 179, 183, 184, 187

under a description 65–72, 79, 81,
89–90, 114

INDEX204

Vesey, Godfrey 86, 172
virtues 27, 36, 41, 44, 47; see

also ethics
Vogler, Candace 38n6, 139n6, 144–5
Von Wright, G. H. 144n7

wanting see desire
Wittgenstein, Ludwig 4, 5, 6, 8, 12,

13–17, 50n3, 56, 58–62, 63–5, 77,

104, 105, 119, 155n1, 164n4,
182n3, 183

Will, act of 27, 40, 47, 130
Warnock, Mary 12
Women in philosophy 11–13
World War II: influence on University

life 11–12; use of atomic weapons 17,
28–31; see also Truman

Wright, Crispin 181n2

Routledge Philosophy Guidebook to Anscombe’s Intention
Contents
Acknowledgements
Introduction
Notes

1 Elizabeth Anscombe and Intention
Anscombe’s times
Oxford philosophy
Anscombe’s faith
Intention’s influence
Notes
Suggested reading

2 Three tasks for Intention
The genesis of Anscombe’s Intention
Modern moral philosophy
Intended and foreseen consequences
Intention with which
An account of action
Notes
Suggested reading

3 ‘The subject under three heads’
The subject introduced under three heads
The ‘connective’ approach
‘We are in the dark about the character of the concept’ (§1)
What is this description for?
‘A tool for the philosophy of action’
The structure of Intention
Notes
Suggested reading

4 Intentional action (§§5–19)
The question ‘Why?’ (§§5–18)
‘It was involuntary’ and reason vs cause (§§5–11)
Known without observation (§8)
Involuntary movements (§§7–11)
Intention, motive, cause, reason (§§10–14)
Voluntary vs intentional actions (§§17–19)
Notes
Suggested reading

5 Intention with which (§§20–40)
Intentional actions, further intentions in acting, intentions for the future (§§20–21)
Doing X in order to do Y (§22)
Doing Y in doing X (§§23–28)
Non observational knowledge? (§§29–33)
Practical reasoning as ‘ordinary reasoning’ (§§33–34)
The idea of logical compulsion (§33)
Wanting (§§34–40)
Notes
Suggested reading

6 The character of the concept of intention (§§42–49)
Form of description of events (§§46–48)
Diagnosing the Cartesian impulse
Practical knowledge (§§28–32, §§45–48)
Notes
Suggested reading

7 Expressions of intention for the future (§§2–3 and §§50–52)
The place of expressions of intention for the future in Intention
Species of prediction (§§2–3)
‘I am going to do it unless…’
Intention and the future
Notes
Suggested reading

Bibliography
Index