Practical Knowledge: Selected Essays

1st Edition Kieran Setiya

Visit to download the full and correct content document:
https://textbookfull.com/product/practical-knowledge-selected-essays-1st-edition-kiera
n-setiya/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Cinema, Culture, Scotland: Selected Essays 1st Edition

Mcarthur

https://textbookfull.com/product/cinema-culture-scotland-
selected-essays-1st-edition-mcarthur/

Selected Contemporary Essays Saumitra Mohan

https://textbookfull.com/product/selected-contemporary-essays-
saumitra-mohan/

From the bottom up selected essays 1st Edition

Greenawalt

https://textbookfull.com/product/from-the-bottom-up-selected-
essays-1st-edition-greenawalt/

Shakespeare on page & stage : selected essays 1st

Edition Wells

https://textbookfull.com/product/shakespeare-on-page-stage-
selected-essays-1st-edition-wells/
The Philosophical Imagination Selected Essays 1st
Edition Richard Moran

https://textbookfull.com/product/the-philosophical-imagination-
selected-essays-1st-edition-richard-moran/

Food and Philosophy Selected Essays 1st Edition Spencer

Wertz

https://textbookfull.com/product/food-and-philosophy-selected-
essays-1st-edition-spencer-wertz/

Consciousness and meaning : selected essays First

Edition Balog

https://textbookfull.com/product/consciousness-and-meaning-
selected-essays-first-edition-balog/

The Work of Music Theory Selected Essays 1st Edition

Thomas Christensen

https://textbookfull.com/product/the-work-of-music-theory-
selected-essays-1st-edition-thomas-christensen/

The Virtues of Freedom: Selected Essays on Kant 1st

Edition Paul Guyer

https://textbookfull.com/product/the-virtues-of-freedom-selected-
essays-on-kant-1st-edition-paul-guyer/
PRACTICAL KNOWLEDGE
 OXFORD MORAL THEORY
Series Editor
David Copp, University of California, Davis

Drawing Morals Having It Both Ways

Essays in Ethical Theory Hybrid Theories and Modern
Thomas Hurka Metaethics
Edited by Guy Fletcher and
Commonsense Consequentialism
Michael Ridge
Wherein Morality Meets Rationality
Douglas W. Portmore Motivational Internalism
Edited by Gunnar Björnsson,
Against Absolute Goodness
Caj Strandberg, Ragnar Francén
Richard Kraut
Olinder, John Eriksson, and Fredrik
The Lewd, the Rude and the Nasty Björklund
Pekka Väyrynen
The Meaning of ‘Ought’
In Praise of Desire Beyond Descriptivism and
Nomy Arpaly and Timothy Expressivism in Metaethics
Schroeder Matthew Chrisman
Confusion of Tongues Practical Knowledge
A Theory of Normative Language Selected Essays
Stephen Finlay Kieran Setiya
The Virtues of Happiness
A Theory of the Good Life
Paul Bloomfield
PRACTICAL
KNOWLEDGE
Selected Essays

Kieran Setiya

1
1
Oxford University Press is a department of the University of Oxford. It furthers
the University’s objective of excellence in research, scholarship, and education
by publishing worldwide. Oxford is a registered trade mark of Oxford University
Press in the UK and certain other countries.

Published in the United States of America by Oxford University Press

198 Madison Avenue, New York, NY 10016, United States of America.

© Oxford University Press 2017

All rights reserved. No part of this publication may be reproduced, stored in

a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by license, or under terms agreed with the appropriate reproduction
rights organization. Inquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above.

You must not circulate this work in any other form

and you must impose this same condition on any acquirer.

Library of Congress Cataloging-in-Publication Data

Names: Setiya, Kieran, 1976– author.
Title: Practical knowledge : selected essays / Kieran Setiya.
Description: New York, NY : Oxford University Press, 2016. | Series: Oxford
moral theory | Includes index.
Identifiers: LCCN 2016007614 | ISBN 9780190462925 (hardcover : alk. paper)
Subjects: LCSH: Act (Philosophy) | Intentionality (Philosophy)
Classification: LCC B105.A35 S47 2016 | DDC 128/.4—dc23 LC
record available at https://lccn.loc.gov/2016007614

9 8 7 6 5 4 3 2 1

Printed by Sheridan Books, Inc., United States of America

CONTENTS

Acknowledgements vii

Introduction 1

PART I : Action

1. Practical Knowledge 39
2. Practical Knowledge Revisited 62
3. Sympathy for the Devil 73
4. Knowledge of Intention 107
5. Knowing How 135
6. Anscombe on Practical Knowledge 156

PART II : Ethics

7. Is Efficiency a Vice? 171

8. Cognitivism about Instrumental Reason 180
9. What Is a Reason to Act? 207
10. Intention, Plans, and Ethical Rationalism 226
11. Akrasia and the Constitution of Agency 253
12. Hume on Practical Reason 272

Index 301
ACKNOWLEDGEMENTS

Many people have helped in the writing of these essays. Specific acknowledge-
ments appear in the footnotes of individual chapters, but I give special thanks
to Arden Ali, David Copp, Peter Ohlin, Nancy Rebecca, Emily Sacharin, and
Andrew Ward for assistance and advice in assembling the book. For permis-
sion to reprint previously published work, I am grateful to John Wiley and
Sons, the University of Illinois, the University of Chicago, Oxford University
Press, Harvard University Press, the Aristotelian Society, and Springer, as
follows:

‘Hume on Practical Reason’, Philosophical Perspectives 18: 365–​89,

© John Wiley and Sons, 2004
‘Is Efficiency a Vice?’ American Philosophical Quarterly 42: 333–​9,
© University of Illinois Press, 2005
‘Cognitivism about Instrumental Reason’, Ethics 117: 649–​73, © 2007 by
The University of Chicago
‘Practical Knowledge’, Ethics 118: 388–​4 09, © 2008 by The University of
Chicago
‘Practical Knowledge Revisited’, Ethics 120: 128–​37, © 2009 by The
University of Chicago
‘Sympathy for the Devil’, Chapter 5, pp. 82–​110, Desire, Practical Reason,
and the Good, edited by Sergio Tenenbaum, © Oxford University
Press, 2010
‘Knowledge of Intention’, Essays on Anscombe’s Intention, edited by
Anton Ford, Jennifer Hornsby, and Frederick Stoutland, eds.,
© Harvard University Press, 2011
‘Knowing How’, Proceedings of the Aristotelian Society 112: 285–​307,
reprinted by courtesy of the Editor of the Aristotelian Society, © 2012
v i i i   
• Acknowledgements

‘What Is a Reason to Act?’ Philosophical Studies 167: 221–​35, © Springer,

2014
‘Intention, Plans, and Ethical Rationalism’, Chapter 4, pp. 56–​82,
Rational and Social Agency: The Philosophy of Michael Bratman, edited
by Manuel Vargas and Gideon Yaffe, © Oxford University Press, 2014
INTRODUCTION

This book has two themes: the nature of intentional action and the
foundations of ethics. What is it to act for reasons and so to act
intentionally? And why are certain facts reasons to act in one way
or another, considerations that count in favour of doing so? Such
reasons fix what we ought to do.
Each question is of interest in itself and some of the essays deal
with one to the exclusion of the other. But the themes are closely
related. In particular, there is an approach to ethics I have called
‘ethical rationalism’, which aims to derive the normative facts—​
what there is reason for us to do—​from the nature of agency or the
will.1 According to the rationalist or ‘constitutivist’, the standards
of practical reason are explained by what it is to act intentionally, or
to have the capacity to do so. In one way or another, action theory
is the basis of ethics.
My relationship to this approach is complicated and it plays
a special role in the essays to come. Unlike some, I think it is
possible to construct a compelling argument from premises in
the philosophy of action to ethical conclusions: from the meta-
physics of agency to the norms of practical reason.2 This strategy
has much wider application than is often assumed. While the
rationalist approach has been associated with Kant, who aimed
to derive the moral law from the idea that we act ‘under the idea
of freedom’, a less ambitious rationalist might derive the norm
of means–​end efficiency from the role of desire in motivation,

1. This terminology is introduced in Setiya 2007.

2. For accounts of this argument, see ‘Murdoch on the Sovereignty of Good’ (Setiya
2013a; available online at http://​quod.lib.umich.edu/​p/​phimp/​3521354.0013.
009/​1), ‘Intention, Plans, and Ethical Rationalism’ (this volume: Ch. 10) and
‘Akrasia and the Constitution of Agency’ (this volume: Ch. 11).
2   
• Introduction

or the irrationality of akrasia from the idea that we act ‘under the guise of
the good’.3
Despite a vivid sense of the power of ethical rationalism, and of its metaphys-
ical and epistemic virtues, I do not in the end accept it. Although the action the-
ory assumed by the ethical rationalist need not be extravagant—​far from it—​I
think the nature of intentional action, and of acting for reasons, is more mini-
mal or impoverished than the argument requires. One purpose of engaging in
the details of action theory, as in Part I of this book, is to support this claim. It is
distinctive of my approach that I do so while defending a conception of agency
that is in certain ways demanding, a conception inspired by remarks on practical
knowledge in Elizabeth Anscombe’s pioneering book, Intention. By ‘practical
knowledge’ Anscombe means our distinctive knowledge of what we are doing
when we are doing something intentionally, and of why we are doing it. One of
my central claims is that we cannot explain such knowledge, which many find
puzzling, without appeal to practical knowledge in a second sense: knowledge
how to do what one intentionally does. ‘Practical knowledge’ can be used in a
third way, for knowledge of practical reason, knowledge that is ethical in the
broadest sense of the term. It is in this sense that Part II is concerned with ethics
and, directly or indirectly, with knowledge of what to do. The title of the book
thus applies, in one way or another, to everything contained in it.
In the rest of this introduction, I sketch in more detail how I think about
the project of action theory, how my conception of practical knowledge has
evolved over the last ten years, and how reflection on agency has implications
for ethics. It may be useful to state in advance, without elaboration, some the-
ses I defend. In Part I:

The idea of practical knowledge—​knowledge of what one is doing or

what one is going to do that does not rest on sufficient prior evidence—​
is central to our understanding of intentional action. The capacity to
act for reasons is the capacity for practical knowledge.
Such knowledge rests on, and is partly explained by, practical knowl-
edge in the second sense, of knowing how.
It does not rest on practical knowledge in the third sense: knowledge
of ethical facts. More generally, in acting for reasons, we do not act
‘under the guise of the good’: we need not represent our action as a

3. The Kantian strategy has been pursued by Christine Korsgaard (1996, 2009); alternatives
are considered in Setiya, ‘Intentions, Plans’ and ‘Akrasia’, and in Reasons without Rationalism
(Setiya 2007: Part Two).
 Introduction ↜渀屮↜渀 • ↜渀屮↜渀 3

good thing to do, or the grounds on which we are acting as normative

reasons that support it.
Practical knowledge is knowledge in intention, where intention
involves, but is not reducible to, belief or partial belief.

And in Part II:

Some of what we regard as practical reason is the application of epis-

temology to beliefs that figure in our intentions. In particular, this is
true of the instrumental principle, that one must intend the necessary
means to one’s ends.
A normative reason is a premise of sound practical thinking. In this
sense, reasons are considerations that move us insofar as we are practi-
cally rational.
There is a valid argument from the metaphysics of agency to the norms
of practical reason, of the sort invoked by ethical rationalists.
But ethical rationalism is false: the nature of agency is too thin to pro-
vide its premises.
The standards of practical reason are standards of ethical virtue, applied
to practical thought.

In relation to the last two theses, the defence offered here is partial: further
arguments appear in Reasons without Rationalism, to which this collection is
at once a sequel and a preface. The essays that follow are independent of that
book, but they deal with related topics. My hope is that reading them will give
a clearer sense of the difficulty, and the urgency, of its project: to make sense of
rational agency and reasons to act outside the context of ethical rationalism.
At the same time, I hope they make progress with some of the most intriguing
puzzles in the philosophy of action, quite apart from their connection with
ethics. It is to those issues that I now turn.

1.╇What Is Action Theory?

Action theory is concerned in the first instance with what it is to act for rea-
sons. It aims to understand the kind of explanation of what someone is doing
that cites the reasons for which she is doing it. Explanations of this sort are
4   
• Introduction

often teleological: ‘A is buying fish in order to cook dinner’. But they also take
non-​teleological forms, as when we state the fact, or putative fact, that is the
reason for which someone acts: ‘A is returning the book on the ground that
he promised to; that is among his reasons for doing it’. When an explanation
of either kind is true, it follows that A is acting intentionally. The converse
implication is less clear: Anscombe disputes it; Donald Davidson responds.4
More important for our purposes is the well-​marked ambiguity of ‘reason’, a
term that appears both in statements of the reason for which someone acts, like
those above, and in statements of the reasons there are for acting in one way or
another, considerations that count pro and con: ‘The fact that his friend is in
need is a reason for A to help’. Philosophers call the latter ‘normative reasons’.
The logic of normative reasons is quite different from that of reasons-​for-​which.
When A is φ-​ing on the ground that p, it follows that A is φ-​ing, and arguably
that A believes that p; it at least doubtful whether it follows that p.5 When the
fact that p is a reason for A to φ, it follows that p, but not that A believes that p or
that she is φ-​ing. Other connections are in dispute. Does it follow, when A is φ-​
ing on the ground that p, that she represents the fact that p as a normative reason
to φ? Some philosophers say yes; I argue that the answer is no.6
What I want to address now is not that question but a more abstract one,
about the aims and ambitions of action theory. In my view, the principal aim
can be stated quite simply. We want to know if the following principles can be
completed without circularity, and if so, how:

To φ intentionally is to φ …
To φ on the ground that p is to φ …

Our attempts at a theory of what it is to act intentionally, or to act on the

ground that p, ought to explain how these phenomena relate to others: why
acting on the ground that p requires the belief that p, assuming it does, how
it involves the agent’s intentions, what intentions are, and so on. These are
among the issues addressed by the essays in Part I.
Unfortunately, we cannot leave the subject here. For the philosophical
treatment of action is often introduced in ways that conflict with mine, ways
that import demands extraneous to action theory, or that leave its object
obscure. Most prominent here is the invocation of Wittgenstein, who asked,

4. Anscombe 1963: 25; Davidson 1963: 6.

5. See Dancy 2000: 132. I expand on this in Setiya 2011: 132–​134.
6. In Setiya, ‘Sympathy for the Devil’ (this volume: Ch. 3) and ‘Akrasia’.
 Introduction      •      5

in the Investigations, ‘What is left over if I subtract the fact that my arm goes
up from the fact that I raise my arm?’ (Wittgenstein 1953: §612). In my view,
this question is seriously misleading.7 It is unhelpful, first, in taking a ‘subtrac-
tive’ form, which suggests an additive theory: what is left over when I sub-
tract the fact that my arm goes up from the fact that I raise my arm is X,
so raising my arm consists in X plus my arm going up. Looking for theories
that have this shape is arbitrarily restrictive. Compare a simple approach on
which I raise my arm just in case my arm goes up because I intend it to. If you
subtract the fact that my arm goes up, what is left is my intention, causing
nothing. You cannot construct an arm-​raising, even on this simple approach,
just by adding intention to my arm’s going up: you need the causal relation.
This defect is superficial; we need not assume that the account of what
it is for me to raise my arm will take a conjunctive form. Instead, we can ask
whether and how it is possible to complete this formula without circularity:

For me to raise my arm is for my arm to go up …

But this, too, is unhelpful, because it is too general. We can ask a similar ques-
tion about the application of any transitive verb. Is there a non-​circular com-
pletion of principles like these?

For the flower to open its petals is for the petals to open …
For the fire to melt the ice is for the ice to melt …

What fills the ellipses may be a further conjunct, a causal explanation, or some-
thing else. The project of spelling it out is not specific to intentional action,
nor is it clear what motivates it. Why think that the application of transitive
verbs can be explained in terms of their intransitive counterparts? And why
suppose that the question is philosophically urgent? Is there some basis for the
primacy of the intransitive? A puzzle in the metaphysics of transitive verbs?
A more radical but more principled approach would aim at a reduction
of dynamic phenomena in general. The contrast between static and dynamic
properties corresponds to the linguistic contrast between verbs that take
progressive or perfective aspect and ones that do not.8 Some verbs have two

7. As Wittgenstein would agree, though for different reasons: ‘When I raise my arm I do not
usually try to raise it’ (Wittgenstein 1953: §622). There is no inner state—​trying, willing,
intending—​whose presence is a condition of raising my arm.
8. A classic treatment is Comrie 1976, though the distinction has philosophical roots; see
Vendler 1957; Kenny 1963: 171–​186. Later discussions include Mourelatos 1978, Graham
6   
• Introduction

forms, one progressive—​‘The floor was shaking’; ‘He was buying a house’—​
the other perfective, indicating completion or the fact that something
happened: ‘The floor shook’; ‘He bought a house’. Others admit no such dis-
tinction: ‘The fruit was red’; ‘She knew everything’. These sentences do not
report a completed act or event, but a state or condition that something was
in. When verbs of the first kind are used in the present tense, they either have
progressive aspect—​‘ The floor is shaking’; ‘He is buying a house’—​or they are
habitual, indicating a repeated or serial action: ‘The floor shakes’; ‘He buys
houses’. Outside of special contexts, like certain forms of narrative, there is no
present perfective. Verbs of the second kind, which admit no distinction of
progressive and perfective aspect, have a non-​habitual use in the simple pres-
ent: ‘The fruit is red’; ‘She knows everything’.
Though it is introduced linguistically, the distinction here is metaphysical.
Some of the things we predicate of objects can be instantiated ‘perfectively’
and in that sense done, while others cannot. There is no standard terminol-
ogy for this distinction. We can use ‘state’ for properties that lack perfective
instantiation. But there is no obvious term for the rest. It is tempting to call
what can be done in the perfective sense an act. But in this sense, acts can be
performed by inanimate objects, like the flower or the fire, that fall outside
the scope of action theory. We might try ‘event’. But there are problems here,
too. What we mean to identify, in contrast with states, are things predicated
of or instantiated by objects, picked out by verbs like ‘shake’ and ‘buy’. ‘Event’
is typically used, instead, for the referents of noun phrases like ‘the shaking
of the floor’, ‘his purchase of a house’. Though there is a close relation here,
events in this sense are not our primary topic.9 We are interested in what it
is for agents to do things, to instantiate properties of certain kinds. Because
I cannot think of a noun to contrast with ‘state’ that is neither misleading
nor arbitrary, I use the adjective ‘dynamic’ for the properties in question.
With this background, we can locate a possible project, of explaining what
it is to instantiate a dynamic property—​to shake, or buy something—​in terms
of states of objects and relations among them. Perhaps there is metaphysical
pressure to think of reality as fundamentally static. But although it may be
more principled than the ‘primacy of the intransitive’, it is clear that this proj-
ect, too, has no essential place in action theory: in an account of what it is to

1980, and Galton 1984. I explore the distinction, and its relation to epistemic agency, in
Setiya 2013b.

9. See Hornsby 1997: 87–​92 on actions as events and as things done.

 Introduction      •      7

act for reasons. It is a project in general metaphysics that action theorists may
or may not embrace.
What, then, is the project of action theory? It is unhelpful to introduce it by
citing the difference between things I do and things that merely happen to me,
as if emphasizing that word is enough to specify our topic. In one sense, opening
the petals is a thing the flower does, and opening is what the petals do: these are
doings in that they are dynamic properties. What is more, it is clear that when
the doctor taps my knee, I kick my leg. That is something I do, not something
that just happens to me, even when I do it by reflex. It is very different when the
doctor lifts my leg to examine it more closely. That merely happens to me: I do
not lift my leg; he does. But this is not the line that action theorists want to
draw. When they distinguish the actions that interest them from mere happen-
ings, kicking my leg by reflex is meant to fall in the second class.
Nor does it help to emphasize the word ‘I’: ‘What is the difference between
things I do and things that merely happen to me?’ In ordinary terms, it is clear
that I kick my leg as a matter of reflex action. The sense in which it wasn’t
really me, if any, is not something of which we have a pre-​theoretic grasp, to
be illuminated by the philosopher, but a fragment of tendentious philosophy,
on which acting for reasons is explained in terms of the agent’s identity. That
might be right, but we should not appeal to it in specifying what we want to
explain. The same is true of ‘identification’, which is a technical term in need
of definition. Once defined, it is a term we may use in giving a philosophical
account of agency, but we need first to locate the target of that account.
To say that our topic is ‘action’, unqualified, is not to make progress, since
the term could apply to the doings of flowers and fires, and to kicking my leg
by reflex. We need to identify a special class of actions in this encompassing
sense. In my view, the right way to do so is the one with which I began. We
want to know if the following principles can be completed without circular-
ity, and if so, how:

To raise my arm intentionally is to raise my arm …

To raise my arm on the ground that p is to raise my arm …

The action theorist may take for granted the existence of dynamic phenom-
ena, like raising my arm, in explaining what it is to do so intentionally, or
for reasons. She need not attempt to answer Wittgenstein’s question in any
form. There is nothing illicit about this. It may turn out, in the course of
understanding intentional action, that we need a reductive account of some
8   
• Introduction

dynamic properties, or all of them. But we should not assume this from the
start. It would be a substantial discovery that action theorists must, or can,
take on such metaphysical ambitions.
Here I agree with the conclusion, if not the argument, of a notorious pas-
sage in Anscombe’s Intention, §19:

[In] describing intentional actions as such, it will be a mistake to look

for the fundamental description of what occurs—​such as the move-
ments of muscles or molecules—​and then think of intention as some-
thing, perhaps very complicated, which qualifies this. The only events
to consider are intentional actions themselves, and to call an action
intentional is to say it is intentional under some description that we
give (or could give) of it. (Anscombe 1963: 29)

Referring to actions ‘under descriptions’ is Anscombe’s way of focusing

attention not on events picked out by noun phrases—​‘Kieran’s raising of his
arm’—​but on properties instantiated by agents. In our terms, her point is that
we should take for granted the description of what an agent is doing when he
is φ-​ing intentionally and investigate how he is doing it. That is what I have
been urging on methodological grounds, not by arguing that it is impossible
to approach the topic with further reductive aims, but that it is unnecessary.
As it happens, Anscombe believes that the search for a non-​circular com-
pletion of our formula, ‘To φ intentionally is to φ … ’, is hopeless. But this
requires a further argument. Despite appearances, the conclusion of §19,
that ‘an action is not called “intentional” in virtue of any extra feature which
exists when it is performed’ (Anscombe 1963: 28) is not the conclusion she
later reports in similar terms, that being the execution of intention is not
‘a mere extra feature of events whose description would otherwise be the
same’ (Anscombe 1963: 88). The conclusion of §19 is that if there is a non-​
circular account of what it is to φ intentionally, it will be in terms of φ-​ing, not
in terms of ‘preintentional movements’ accompanied by some special feature,
I. The conclusion of the later argument, in §§46–​48, is that there is no such
account. It is addressed to a view that tempts us when the description of what
someone is doing could apply to behaviour that is intentional or not: ‘What is
the difference between merely φ-​ing and φ-​ing intentionally, or on the ground
that p?’ What is the extra feature of this event, whose description would oth-
erwise be the same?
Anscombe’s reason for doubting that there is any such feature turns on the
involvement of practical knowledge, and knowledge of reasons, in intentional
 Introduction ↜渀屮↜渀 • ↜渀屮↜渀 9

action.10 Although I believe that there is real insight in Anscombe’s appeal to

such knowledge, I do not agree with her about its nature or scope, and I do
not believe that it precludes a non-╉circular theory of intentional action. In the
next two sections, I say more about my commitments in this area, how they
have shifted over time, and how they are traced through the essays in Part I.

2.╇Knowledge in Intention
Begin with a simple claim:

(K)╇When A is φ-╉ing intentionally, A knows that he is φ-╉ing.

In Thought and Action, Stuart Hampshire wrote that ‘if a man is doing some-
thing without knowing that he is doing it, then it must be true that he is not
doing it intentionally’; ‘doing something […] intentionally […] entails
knowing what one is doing’ (Hampshire 1959: 95, 102). Two years earlier,
Anscombe published the first edition of Intention, holding that intentional
action is that ‘to which a certain sense of the question “Why?” is given appli-
cation; the sense is of course that in which the answer, if positive, gives a rea-
son for acting’; this question is ‘refused application by the answer: “I was not
aware I was doing that”â•›’ (Anscombe 1963: 9, 11). Anscombe went on to clas-
sify knowledge of our intentional actions as ‘knowledge without observation’,
meaning not only that it is not a matter of perceiving what we are doing, but
that it is not inferred from other facts we know about ourselves.11 If K is a nec-
essary truth, there is pressure to reject an inferential model of the knowledge
it ascribes: what could prevent me from forming and executing the intention
to φ, thus φ-╉ing intentionally, without making the relevant inference?12
Much of the interest and the controversy around such knowledge is epis-
temic. How is it possible, critics ask, for me to know what I am doing except
on the basis of sufficient prior evidence?13 Difficulties here might lead us to
question Anscombe’s insight, or Hampshire’s, in stating K. But there are more
mundane objections to their approach. For there are apparently obvious, com-
pelling exceptions to K. Perhaps the most notorious is due to Davidson, who

10. See ‘Anscombe on Practical Knowledge’ (this volume: Ch. 6).

11. This is made explicit at Anscombe 1963: 50.
12. I make this argument at greater length in ‘Practical Knowledge’ (this volume: Ch. 1).
13. See, especially, Grice 1971; Langton 2004; Paul 2009.
1 0   
• Introduction

imagines a carbon-​copier trying to make ten copies at once.14 If he is succeed-

ing, Davidson suggests, the carbon-​copier is making ten copies intentionally.
But he may not know, or even believe, that the copies are going through so
many times. Davidson concludes: ‘It is a mistake to suppose that if an agent is
doing something intentionally, he must know that he is doing it’ (Davidson
1978: 91).
In his first discussion of this problem, Davidson concedes that, while the
carbon-​copier may not know that he is making ten copies, what he is doing ‘is
known to him under some description’ (Davidson 1971: 50). Inspired by this,
I once proposed a replacement for K:

(B) When A is φ-​ing intentionally, A believes that he is φ-​ing, or else he is

φ-​ing by doing other things, in which he does believe.15

Thus, the carbon-​copier is making ten copies by pressing hard on the carbon-​
paper, he is pressing hard intentionally, and he believes that he is doing so.
Since beliefs of this kind are necessarily present in intentional action, there
is the same pressure to deny that they are formed by inference. The revision
preserves what is most interesting, and most puzzling, in K.
But it does not go far enough. Suppose, for instance, that I have been
recently paralysed, and I attempt to clench my fist, under anaesthetic, behind
my back. If I succeed, I am clenching my fist intentionally, but I do not know,
or believe, that I am doing so. Nor do I clench my fist by taking further means
that meet this condition: means I know, or believe, I am taking. This prompts
a further revision:

(C) When A is φ-​ing intentionally, A is more confident that he is φ-​ing than

he would otherwise be.16

This condition is met in recent paralysis. But the examples do not end. Thus
Sarah Paul imagines a case in which I am less confident that I am φ-​ing when
I do so intentionally than otherwise. She appeals to the neurological disorder
‘Alien Hand Syndrome’, in which I am disposed to perform elaborate actions

14. Davidson 1971: 50, 1978: 91–​92.

15. Adapted from ‘Belief ’, in Setiya 2007: 26; see also ‘Practical Knowledge’ and ‘Practical
Knowledge Revisited’ (this volume: Ch. 2).
16. This formula adapts and simplifies the treatment in ‘Practical Knowledge’; see also Pears
1985: 79–​82.
 Introduction      •      11

unintentionally (Paul 2009). But we need not go so far. If I try to breathe

steadily, my confidence that I am doing so may be less than when I trust my
breathing to my autonomic system.17 Still, if I am breathing steadily, I am
doing so intentionally. Whether C is refuted by this case depends on the rel-
evant contrast, the situation to which we look in determining how confident
I would otherwise be. Is it a situation in which I do not try to breathe steadily
and my autonomic system kicks in, or one that eliminates my attempt without
replacing it with another potential cause? Is there a principled way to decide?
One might respond to these problems by revisiting K.18 When we reflect
on how little is involved in an action’s being in progress—​that one can be
φ-​ing but never φ, or even get close to doing so—​can we credit A with knowl-
edge that he is φ-​ing even in the cases described above? If it is enough to be
in the process of φ-​ing that one take means by which one intends to φ, the
carbon-​copier can know that he is in the process of making ten copies, even
if he does not know that he will—​as you might know that you are building a
house that in fact you will never complete. If it is enough that one intend to
φ, I can know I am in the process of clenching my fist even when I doubt that
it will move.
I do not find these options credible. Anscombe is right to reject the ‘false
avenue of escape’ (from doubts about practical knowledge) on which ‘I
really “do” in the intentional sense whatever I think I am doing’ (Anscombe
1963: 52). If I intend to clench my fist and it fails to move, an observer would
flatly deny that I am clenching it. Nor is there some hidden ambiguity here,
‘two objects of knowledge’, practical and theoretical, picked out by the same
words (Anscombe 1963: 57). The proposition I know when I know that I am
clenching my fist because I am doing so intentionally is the one an observer
knows when he sees it begin to move; and it is one whose truth I cannot know
when I believe I am still paralysed.
A better way to save the idea of practical knowledge is not to weaken its
content but to turn from instances of such knowledge to the capacities that
afford it.19 The capacity to act for reasons, and so to act intentionally, is a
capacity for knowledge of what one is doing and why. Like other cognitive
capacities, this one can be exercised imperfectly, issuing in mere belief instead

17. See ‘Practical Knowledge Revisited.’

18. Perhaps inspired by Falvey 2000 or Thompson 2008: Part One. I consider this response
more fully in ‘Anscombe on Practical Knowledge’; see also the notes on Thompson below.
19. As in ‘Knowledge of Intention’ (this volume: Ch. 4) and ‘Knowing How’ (this
volume: Ch. 5).
1 2   
• Introduction

of knowledge, or in partial belief, a degree of confidence that p. It is this capac-

ity that action theory needs to comprehend. In my view, it should do so by
picturing intention as a cognitive state, one that amounts to knowledge, or
belief, in many cases, but which is consistent with serious doubt.
This way of putting things prompts several questions. First, it is one thing
to insulate the alleged insight about intentional action and knowing what one
is doing from obvious counterexamples. It is another to argue for its truth.
Why think of the capacity for intentional action as a cognitive capacity at all?
Second, there is still the question how such knowledge is possible, how we can
know what we are doing except on the basis of sufficient prior evidence. And
third, there is the need to say more about the kind of capacity that constitutes
the will: what makes this capacity practical as well as epistemic? Let me take
these points in turn.
As to the first, there are several reasons. For one, the default expression of
my intention in acting takes the form of an assertion, ‘I am φ-​ing’—​or in the
case of prospective intention, ‘I am going to φ’—​that looks like the expression
of belief. This comes out in teleological explanations of action that take what
Michael Thompson calls a ‘naïve’ form.20 Asked, ‘Why are you boiling water?’,
I reply that I am making tea, since I am boiling water in order to make tea.
This makes sense if the assertion, ‘I am making tea’, expresses the intention
with which I am boiling water, not some further mental state. In addition,
there are indirect arguments for the cognitive conception of the will, some
of which appear below. There is an argument from the nature of instrumen-
tal reason, an argument from self-​knowledge of intention, and an argument
from the epistemology of knowing how.21 Taken together, these arguments
convince me that the cognitive conception is right.
On the second question, I explain the possibility of practical knowledge—​
knowledge in intention of what one is doing or what one is going to do—​by
appeal to knowledge how. This view occupies an intermediate place between
two extremes. According to the practical knowledge sceptic, one is never
justified in forming a belief without sufficient prior evidence, as one does in
forming an intention on the cognitive account.22 On the other, permissive
extreme, one is justified in forming a belief so long as one knows that the

20. In Thompson 2008: Part One.

21. See, respectively, ‘Cognitivism about Instrumental Reason’ (this volume: Ch. 8),
‘Knowledge of Intention’, and ‘Knowing How’.
22. See the authors cited in note 13.
 Introduction      •      13

belief will be true, and supported by evidence, once formed.23 I agree with the
sceptic in finding this insufficient. The force of the complaint is clear in con-
nection with knowledge. When you come to know that p, the truth of your
belief is credited to you: it turns on dispositions or capacities whose opera-
tion makes the truth of that belief no accident. This condition is absent from
the permissive view, which is thus too liberal. But the sceptic’s position is too
extreme. For the condition can be met in at least two ways: by forming beliefs
on the basis of sufficient prior evidence, or by forming beliefs one is disposed
to make true. One is justified in forming the intention to φ, with its correlative
belief, only if one knows how to φ, where knowing how to φ is a disposition
to execute that intention whose operation makes the truth of one’s belief no
accident. Since knowledge how comes by degree, the degree of reliability in
one’s disposition, we should expect a similar gradation in the strength of the
beliefs one is entitled to form.
The details of this sketch are spelled out in Part I, though it leaves some
questions open. One, in particular, I would like to close. In coming to know
that one is φ-​ing, or that one is going to φ, where this is knowledge in inten-
tion, must one know, or be in a position to know, that one is able to φ? I think
the answer must be yes. We would otherwise license an illicit form of ‘boot-
strapping’ in which one decides to φ, comes to know that one is φ-​ing on the
basis of knowing how, and then infers, without evidence, that one is able to φ.
What prevents such bootstrapping is the demand that one know, or be
in a position to know, that one is able to φ. But this demand is potentially
problematic.24 In some cases, one has empirical knowledge that one is able to
φ, as for instance on the basis of past attempts. But not in every case. What
happens when one is not in a position to know in advance that one is able to
φ? How can it be rational to attempt a new intentional action—​say, riding a
bike—​for the very first time, if my intention in acting involves the belief that
I am doing it? In some cases, the answer may be that I do not intend to φ but
to take some possible means: to push the pedals, hold the handlebars, and so
on. But this is not a general solution. For in other cases, it is clear that I intend
to φ even though I am not in a position to know that I can. (I do not merely
intend to try, since my intention is not fulfilled if I try but fail.) This must be
true when I learn new basic actions, not performed by taking further means.

23. See Harman 1976: 164n8; Velleman 1989: 56–​6 4.

24. The problem is raised, but not resolved, in ‘Practical Knowledge Revisited’ (this vol-
ume: 68). As I note in that discussion, the issue here is parallel to problems of bootstrapping
and ‘easy knowledge’ that appear in other domains.
1 4   
• Introduction

The key is to remind ourselves that belief comes by degree. I may not
know, or believe, that I am acting as I intend, though I have some confi-
dence that I am. Likewise, while I may not be in a position to know that
I am able to φ when I begin to learn, I must already be entitled to some
degree of confidence that I am able to φ. This is less peculiar than it seems.
What I am in a position to know is, in effect, that I might be able to φ, that
there is some prospect of success. It is this minimal entitlement, together
with knowledge how, that makes it rational to try. As to where the entitle-
ment originates, there are two options. One is that, in order to be rational
in attempting to φ, one must have at least a shred of empirical evidence
that one can do it, perhaps from one’s success at related activities, or one’s
general ability to control to one’s body. The other is that one’s confidence
is justified a priori. Either way, the possibility of learning how can be made
consistent with a cognitive theory of intention if, and only if, we allow for
partial belief.
The final question raised above is how the capacity for practical knowl-
edge differs from other cognitive capacities. What makes it distinctively prac-
tical? Part of the answer is contained in my remarks on knowing how. When
one knows that p, it is not an accident that one’s belief is true. In the case of
theoretical knowledge, this condition is met by the origin of one’s belief, its
being formed by a method whose reliability is not an accident, as by appeal
to sufficient prior evidence. In the case of practical knowledge, the condition
is satisfied in part by one’s disposition to act on the belief that figures in one’s
intention, a disposition that constitutes knowledge how.
This contrast—​between two ways in which knowledge can be secured, prac-
tical and theoretical—​is reminiscent of Anscombe on mistakes of performance
and of judgement.25 But it is quite different. Anscombe’s point is not about the
ground of non-​accidental truth but about the locus of error. Anscombe distin-
guishes the standard of mistakes imposed by intention from the one applicable
to belief. When you believe p and p is false, your belief is mistaken. That is a
mistake of judgement. For Anscombe, there is no mistake of judgement when
you are not acting as you intend, even though you think you are φ-​ing and the

25. Anscombe 1963: §32. This section is often cited as the source of a metaphor, that attitudes
can be distinguished by their ‘direction of fit’, with belief on one side, and desire on the other.
These states relate to the world in opposing ways, beliefs being meant to fit the world, desires
to make the world fit them. As Kim Frost has argued, this is both a misreading of Anscombe,
whose discussion is not metaphorical or about mental states in general, and dubious in itself;
see Frost 2014.
 Introduction      •      15

proposition that you are φ-​ing is false.26 There may be a mistake of judgement in
the background, when your failure rests on a false belief about means to ends;
but that is a separate matter. Moreover, when ‘a man is simply not doing what
he says’, in that his failure does not rest on false beliefs, ‘the mistake is not one of
judgment but of performance’ (Anscombe 1963: 57).
The condition just described is sufficient for performance error. Is it also
necessary? Anscombe does not say. But she seems right to insist that one’s
intention in acting sets a standard for what one does, in relation to which one
can make mistakes. In this respect, the capacity for practical knowledge dif-
fers sharply from other cognitive capacities, a difference we can add to the one
described above. It is much less clear that Anscombe is right on the negative
point, that there is no mistake of judgement when this capacity misfires and
one is not acting as one intends. On the more natural view, intention involves
belief, or partial belief, and is subject to the same condition of error. When
I think I am pressing button A, because that is what I intend, but I am press-
ing button B, I mistakenly believe that I am pressing button A, and I make
a mistake in pressing button B. I make mistakes of judgement and perfor-
mance. In support of this we can cite the fact that knowledge implies belief,
so that intention involves belief when one has knowledge in intention, and
the fact that one can make inferences from practical knowledge, or would-​be
knowledge, in just the way one does from other beliefs.27
At the same time, intention is distinctive not just because its standing as
knowledge is secured in a distinctive way, and because it sets a standard for
mistakes of performance, but because it is in the nature of intention to moti-
vate action. Intending involves the kind of wanting whose ‘primitive sign’, in
Anscombe’s words, is ‘trying to get’.28 What this means is, roughly, that when

26. Anscombe is, if anything, more explicit in the case of prospective intention: ‘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.
But why should we not say: this only shows that there are other ways of saying what is not true,
besides lying and being mistaken’ (Anscombe 1963: 4).
27. It is a good question why Anscombe denies that there is a mistake of judgement when I am
not acting as I intend. Perhaps she doubts that it would be rational to form an attitude to p that
is in error when p is false unless one has sufficient evidence that p is true. That principle would
count against the forming of intentions, if intention involves belief. But I do not see the force
of the principle, once we accept the reality of practical knowledge.
28. Anscombe 1963: 68. To think of intention as both cognitive and motivating or ‘desire-​like’
(in the terminology of Setiya 2007) is to avoid the problem of Parfit’s insomniac (discussed
in Harman 1976: §III), whose belief that he will stay awake is self-​fulfilling. No matter what
we add to the content of this belief—​perhaps he believes that he will stay awake because he so
believes—​it does not count as the intention to stay awake.
1 6   
• Introduction

one intends to φ and one can do so directly, without taking further means, one
is disposed to φ in execution of one’s intention; if one is capable of φ-​ing, but
only indirectly, by taking further means, one is disposed to intend the appar-
ent means because one intends to φ. These facts about the ‘functional role’ of
intention might explain why it sets a standard for performance error, as a kind
of malfunction. This standard is clearly violated when one is simply not acting
as one intends, pressing B when one intends to press A. It is less clear whether
it is violated when one’s failure to act as one intends derives from false beliefs
about means, a point reflected in Anscombe’s account.
When I execute my intention, that explains what I am doing: I am
φ-​ing because I so intend. Is this a causal explanation? In an essay written after
Intention, Anscombe emphatically says no: it is a ‘mistake […] to think that
the relation of being done in execution of a certain intention, or being done
intentionally, is a causal relation between act and intention’ (Anscombe
1983: 95). But the issue is obscure. Anscombe assumes that the relata of
causation are states, where ‘a state is supposed to be something holding of
its subject here and now, or over a period of time, without reference to any-
thing outside that of which it holds or the time at which it holds’ (Anscombe
1983: 99). Intending is not a state in this sense, since whether I intend to
visit the bank depends not only on what is true of me, here and now, but on
whether I am acquainted with banks, that is, on facts about my past environ-
ment. An intrinsic duplicate of me in a world without banks could not intend
to visit one. The proper response to this observation is not to deny that inten-
tions are causes, but to resist the view that causation and causal explanation
can only advert to ‘narrow’ or ‘local’ states, in Anscombe’s sense.29
The best way to make sense of the explanation of action by intention is to
think about the manifestation of dispositions or the activation of powers.30
The glass breaks when struck because it is fragile, in that it is disposed to break
when struck: it manifests that disposition. The radioactive isotope decays
because it is disposed to do so: its disposition to decay is realized. In a similar
way, intentions interact with knowledge how, a dispositional state. To mani-
fest know-​how is to execute one’s intention in action; to act intentionally is

29. See Yablo 1997.

30. Here I agree with Hyman 2013. We differ in that he treats desire, and so intention, as a
disposition to act. I think he omits the pivotal role of knowing how. One can intend to φ but
not be disposed to φ because one does not know how to φ. And the disposition one mani-
fests in acting intentionally is a disposition to execute one’s intention in action that constitutes
knowledge how.
 Introduction      •      17

to manifest know-​how.31 If explaining something as the manifestation of a

disposition is causal explanation, so is the explanation of action by intention
and knowledge how.
In each case, there is room for ‘causal deviance’. A fragile glass can break
when struck in a way that does not manifest its disposition to do so: sup-
pose it is attached to an explosive device, triggered by contact, that would
destroy even a sturdy object. When a glass manifests its disposition to break
when struck, its breaking must be caused by being struck ‘in the right way’.
Similarly, I may blink because I intend to do so without manifesting knowl-
edge how to blink: suppose my intention makes me nervous and I am caused
to blink involuntarily.32 Then I do not blink intentionally or in execution of
my intention. For that to happen, my intention must explain my blinking ‘in
the right way’. What counts as the right way depends on the sort of disposi-
tion involved. In connection with knowing how, it is useful to distinguish
two sorts of deviance: basic and non-​basic.33 A basic intentional action is one
that is not performed by means of some other intentional action. Here the
crucial concept is that of guidance: intention and know-​how not only cause
but guide one’s behaviour. Understanding what it is to manifest know-​how is
understanding the way in which it guides intentional action, correcting for
perceived divergence from what the agent intends. When one performs an
intentional action by means of others, to manifest knowledge how is to act in
accordance with one’s plan; those who know how are disposed to formulate
suitable plans and execute them in action.
There is one more source of resistance to intention as cause of intentional
action, which is Aristotle’s claim that action itself is the conclusion of prac-
tical reasoning. Those who find this claim attractive recoil from a picture on
which practical thought terminates with intention, a mental state distinct
from, and productive of, intentional action. This picture is vividly expressed
by John Broome:

[An] action—​at least a physical one—​requires more than reasoning

ability; it requires physical ability too. Intending to act is as close to

31. Since know-​how comes by degree, so does knowledge in intention: I am entitled to believe
that I am φ-​ing, when that is what I intend, only so far as I know how to φ and am entitled to
be confident of my ability.
32. The case is adapted from Davidson 1973: 79.
33. See Davidson 1973: 78–​79; Setiya 2007: 31–​32.
1 8   
• Introduction

acting as reasoning alone can get us, so we should take practical reason-
ing to be reasoning that concludes in an intention. (Broome 2002: 83)

What is at stake in Aristotle’s claim? Presumably, it is agreed on all sides that

there is such a thing as acting for a reason: reasons can attach to and inform
what one is doing, not just what one intends to do. What is more, it is not
just physical ability but also know-​how that informs the causation of action
by intention. How could the relation of reason to action be closer than this?
There are two ways. First, there is Michael Thompson’s idea that intend-
ing to φ is a matter of being embarked on intentional action.34 If one intends
to φ, one is already in progress, though perhaps at an early or liminal stage.
It follows that, if practical thought can reach as far as the intention to φ, it
can reach as far as φ-​ing itself. There is more to say on behalf of this view; but
I am sceptical about it.35 One can intend to φ without even incipiently φ-​ing.
Suppose I intend to perform a basic intentional action that I don’t know how
to perform. Am I bound to be in progress towards doing it? Alternatively,
suppose I intend to do the impossible. I intend to be walking home by the
shortest route but I have taken a serious wrong turn. Is there any sense in
which I am in progress to walking home by the shortest route? There is a fur-
ther difficulty for ‘intending as doing’ which turns on the logical complexity
of objects of intention. I can intend not only to walk home but also to walk
home if there are no taxis, not to walk home, to press button A or button
B, and so on. If intention is an attitude, we can put such complexity in its
object. If intending is being in progress, there must be intentional actions cor-
responding to arbitrarily complex concatenations of intentional actions; but
it is quite unclear what these could be.
The second way to close the gap between intention and action is to go
disjunctive.36 Intention is not a factor or constituent of intentional action that
might cause its execution but a condition that takes two forms: acting inten-
tionally and mere intending. In the cases described above, in which I lack
know-​how or it is impossible to φ, I merely intend to act. When I act inten-
tionally, I do not merely intend to act, not because a mere intention is effec-
tive, but because I am not in that state at all. Practical thought terminates with

34. Thompson 2008: Part One.

35. For more extensive discussion, see my entry on intention in the Stanford Encyclopedia of
Philosophy, online at plato.stanford.edu/​entries/​intention, section 1; and for influential resis-
tance, on the basis of ‘pure intending’, Davidson 1978.
36. As in Davidson 1978: 99; Rödl 2007: Ch. 1.
 Introduction      •      19

intention, that is to say, with mere intending or intentional action. In the lat-
ter case, Aristotle is right: the conclusion of practical reasoning is action itself.
One challenge for any disjunctive view is to say what unifies the disjuncts
that we gather by a single name. There is a particular puzzle here in that the
alleged disjuncts of intention seem to be of different metaphysical kinds: act-
ing intentionally is dynamic, mere intending is static. What brings them
together? The natural response, in the present context, is to identify inten-
tional action with practical knowledge and to deny that knowledge reduces to
belief.37 Instead of being a constituent of knowledge, along with other factors,
belief takes two forms: knowing that p and merely believing that p, the first
of which is not explained in terms of the second. Likewise, intention is not a
constituent of intentional action but a condition that takes two forms: practi-
cal knowledge of what one is doing and merely intending to act, the first of
which is not explained in terms of the second. When I have knowledge in
intention that I am φ-​ing, it follows that I am φ-​ing: knowledge entails and
does not cause intentional action. Nor is it partly composed of intention, or
practical belief, which could be cited as a cause.
A question for advocates of this approach is how they can distinguish
practical knowledge from other forms of cognition without appeal to causal
conceptions of intending and knowing how. But there is a more basic prob-
lem, if my arguments are right, namely that the contrast between practical
knowledge and mere belief does not line up with the contrast between inten-
tional action and mere intending. In some cases of partial belief, like that of
recent paralysis, one executes one’s intention and so acts intentionally, in the
absence of practical knowledge. If intention is a cause of action here—​we
cannot complain that action is entailed, not caused, by intention as practical
knowledge—​why not elsewhere?
The upshot is a theory of intention as a mental state that involves both
desire and belief or partial belief. The content of intention in action is the
proposition that one is φ-​ing; the content of prospective intention is the prop-
osition that one is going to φ. Intention motivates action by way of knowl-
edge how, and sets a standard for mistakes of performance in what one does.
The same state is present when one fails to act as one intends. When one acts
intentionally, intending is no less a cause of action than being struck is a cause
of breaking in the fragile glass. In each case, a causal power is manifested: a

37. A theme of Williamson 2000. I mention this view briefly, and agnostically, in a footnote
to ‘Knowing How’ (this volume: 151n26). Here I reject it, at least in application to practical
knowledge and intention as the cause of action.
2 0 ╇↜
• ↜ ╇ I n t r o d u c t i o n

disposition to break when struck, or to execute one’s intention in action. To

act intentionally is to manifest knowledge how.

3.╇Acting for Reasons

How does this theory of intentional action account for Anscombe’s question
‘Why?’ How does it explain what we do not just intentionally but for reasons?
In Anscombe’s book, these topics are inseparable. Intentional action is that to
which the question ‘Why?’ is given application, in the special sense that asks
for reasons; this question fails to apply when the agent does not know what
she is doing or knows it only on the basis of sufficient prior evidence. In effect,
the theory sketched above and developed in Part I builds an account of inten-
tional action on the second condition: on the idea of knowledge in intention
and its relation to knowing how. It treats acting for reasons as a special though
pervasive case of this phenomenon.
What is involved in acting for a reason? Many philosophers hold that, in
acting on the ground that p, one takes the fact that p as a normative reason
for what one is doing. One must regard the fact that p as a consideration that
counts in favour of one’s action. One of the main contentions of my work on
this topic is that such philosophers are wrong: it is not a condition of acting
for a reason that one represent the ground on which one is acting as a norma-
tive reason to act. Nor does it help to weaken the content of the normative
proposition (from a claim about reasons to one of approximate rationality)
or the attitude one takes to this proposition (from belief to mere seeming).
As I argue in ‘Sympathy’ and ‘Akrasia’, these conditions are equally flawed. In
terminology that is now standard, we do not act intentionally, or for reasons,
‘under the guise of the good’.
At the same time, there is a sense in which, in acting for a reason, one takes
a consideration as one’s reason for acting. The sense is not normative, nor is
the attitude in question mere belief. It is rather that, as one intends to φ in φ-╉
ing intentionally, so one intends to act for the reasons for which one acts. We
decide not only what to do but why.38 Since reasons figure in the content of
our intentions, and intention is a cognitive state, acting for a reason involves
belief, or partial belief, about the explanation of one’s action. That is why we
can respond to Anscombe’s question when we are acting for reasons. If all
goes well, we know what we are doing intentionally and why we are doing it.

38. For recent versions of this claim, see Wallace 1999: 239–╉242; Searle 2001: 16; Setiya
2007: 39–╉49. It draws on the Kantian idea that we will the maxims of our actions.
Another random document with
no related content on Scribd:
 Madidu Elaui Agola Badjehun Karikari
(present Chief) (dead) (dead)

Assalm El Musa Mursa Djamarata Imuhadjil 1 son (?)

i Mekki

ALIMSAR

Durrata Azuhur Fihirun 3 other sons (?)

(dead)

Aneirum. 2 other 2 sons 1 son

sons (?) (?)

Here too is a list of the tribes making up the Awellimiden

Confederation, with the names of their present chiefs.

THE NOBLE OR IHAGGAREN TRIBES.

Kel Kumeden—Chief Madidu. Kel Tekeniuen—Chief Burhan.
Kel Ahara—Chief El Yasan. Kel Takabut—Chief Aluania.
Kel Tedjiuane—Chief Arreian. Teradabeben—Chief Sidauat.
Iderragagen. Tenguereguedeche—Chief Warigoru.
Tarkaitamut. Tademeket—Chief Yunès.
Tahabanat. Idalbabu—Chief Ihuar.
Ibehauen—Chief Sar’adu. Ahianallan.
Ifoghas—Chief Waruziga.
Ihegaren—Chief El Auedech.

SERFS OR IMRADS.
Kel Gossi—Chief Ur illies. Tar’ahil—Chief Ekerech.
Irreganaten—Chief Ur orda. Ikairiraen—Chief Ezemek.
Iueraruarar’en—Chief Mahamud. Erkaten—Chief Elanusi.
Imideddar’en—Chief Huberzan. Ikawellaten—Chief Ibunafan.
Ibongitan—Chief Allabi. Ihaiauen—Chief Abba.
 Tafagagat—Chief Karrabau. Kel R’ezafan—Chief Amachecha.

To these tribes making up the actual Confederation must be

added the following, who were brought into it by force, and have long
since submitted with a good grace to be under the protection of the
Awellimiden:
Wadalen—Chief Niugi. Eratafan—Chief Yoba.
Cheibatan—Chief Rafiek. Ibendasan.
Logomaten—Chief Bokar Wandieïdu. Ahiananurde—Chief Amadida.
Tabotan—Chief Muley.

Subject to each of these last-named tribes are imrads, but I only

know the name of one of their tribes, that of the Ekono, vassals of
the Wadalen.
In addition to their predatory excursions the Tuaregs on the right
and left bank of the Niger make two annual migrations, the time of
which is generally the same.
During the dry season, from December to May, the higher districts
are sterile and dry, the ponds and wells empty of water. Then the
Tuaregs move down to the river-banks and their flocks and herds
graze on the coarse weeds which line them. To avoid the sickness
amongst the camels which results from eating damp food, and to
which I alluded in speaking of Timbuktu, they generally leave them a
little further inland. It is at this time that the negroes pay their tribute
of maize and tobacco, and it is also during this same season that
warlike expeditions are generally undertaken.
For the rest of the year the rain pours down in torrents in the
riverside districts, and although its fall is not so constant or so heavy
in the higher lands, they too are fertilized by the filling up of the
ponds and the wells, many of which even overflow.
Then the nomad tribes go back again to their old haunts, and
settle down for the winter in their camps about the wâdies,
resembling those of Algeria, which begin near Gao.
These wâdies are such very characteristic features of Central
Africa, that a description of one of them may be useful. The word
wâdy means the channel of a watercourse which is dry except in the
rainy season, but there is water in the upper portion of that of Gao in
every season. Its source is far away in the north, and it seems to be
identical with the Igharghar of the south, alluded to by Duveyrier, the
Astapus of the ancients, which comes down from the Atakor or
Ahaggar.
This would confirm Barth’s suggestion, that the marshy
depressions which debouch on the Ngiti Sokoto do not extend
beyond the district of Air.
My own opinion is that the Gao Wâdy, before it became choked
up with sand, was a tributary of the Niger when the course of that
river was far more rapid than it is now.
An examination of its banks does in fact lead to the conclusion,
that nearly if not quite all along them a line of cliffs, eroded by the
action of water, marks what was once the bed of part of the old
Niger. In their annual migration the Awellimiden go up as far as the
districts near Air, where they come in contact with their enemies the
Kel Gheres. Probably competition for the ownership of the
pasturages yielding food in the dry season, was the original cause of
the feud between the two races, which dates from centuries ago.
The tribes from the left bank of the Niger also move into the kind
of islet formed by the bend of the river, advancing to near Dori,
where they find a series of ponds and lakes known as Oursi Beli,
etc., an idea of which I have tried to give in the map accompanying
this volume, but I do not know how far I have succeeded.
There are many very curious and interesting hydrographical
problems connected with this bend of the Niger reserved for the
future explorer to solve.
Well, what do my readers think of the Tuaregs after the picture I
have endeavoured to give of them? I certainly have not represented
them as saints, living in a kind of Utopia, where all is well, where the
men have no vices and the women no faults.
You will perhaps, however, agree with me that they have very
decided characters, and many fine qualities, if also many defects.
Their intelligence is certainly great, making it well worth while to try
and win them to a better mode of life, and one more conducive to the
comfort of their neighbours.
I do not of course fail to recognize what hard work it is to row
against the current or to contend against pre-conceived ideas. It is
always difficult, and sometimes dangerous.
In 1859 a young Frenchman, not more than twenty years old at
the most, disembarked at Constantine. He spent three years
travelling about the Algerian Sahara, and under the powerful
protection of the Emir Ikhenukhen, chief of the Azguers, he lived for
more than a year amongst the Tuaregs.
After his return an expedition was sent out by the Governor of
Algeria, and the treaty of Rhadamès was signed.
Then, in accordance with the traditional French policy in matters
colonial, instead of profiting by the results already acquired,
absolutely nothing further was done. Duveyrier described the
Tuaregs as he had found them, just as I have tried to do; he spoke
quite frankly of their faults as well as of their virtues, and insisted on
the possibility of treating with them on favourable terms. He might
well do so, for he had already succeeded in that direction himself.
When twenty years later Flatters was assassinated, Duveyrier
was accused of mendacious optimism, and every one was ready to
cast a stone at him.
As a matter of fact, however, Flatters was killed by the Hoggars,
and Duveyrier had mentioned that they were living in a state of
anarchy, which seemed likely to get worse and worse rather than to
improve. Flatters insisted on going through their territory, although
the Amrar had told him he could not protect him. Now Duveyrier had
made a special point of never going into any district without first
securing an efficacious safe-conduct, yet in spite of all this he is
made responsible for the disaster.
A fitting epilogue ensued, for Duveyrier, disquieted at the
accusations brought against him, weakened by fever contracted in
his journey, and cut to the heart by the ingratitude of his fellow-
countrymen, committed suicide by shooting himself with his revolver,
in the hope perchance of finding the justice denied him here in
another world, if there be indeed such a thing as justice anywhere.
The English would have made him
a peer, and put up statues in his
honour; the ignorance of the French, I
will not use a harsher word, drove him
to commit suicide.
The example is certainly not
encouraging to us later explorers.
I should have been more likely to
win applause if I had pictured the
TUAREGS.
Tuaregs as irreclaimable savages,
relating a thousand entanglements
with them, such as imaginary conflicts with their armed bands, where
my own presence of mind and the courage of my party saved the
expedition from massacre.
I have preferred in the interests of my country to tell the truth, the
whole truth, and nothing but the truth.
Even as I write these words, I hear of the death of two young
officers and their men, who were killed near Timbuktu in a fight with
a Hoggar razzi. The Hoggars again!
This does but confirm what I insisted on when I was at Timbuktu,
that we shall never succeed in getting en rapport with the nomad
tribes except with the aid of those tribes themselves.
We must first subjugate certain tribes, and then form from
amongst them auxiliary levies, or, as the natives call them, maghzen,
which will aid us, at a minimum cost to ourselves, to establish French
influence over the Tuaregs.
Amongst the tribes who would best lend themselves to this
purpose, I place the Awellimiden in the very first rank, and they are
the hereditary enemies of the Hoggars. Or perhaps I should rather
have said, if we wish to bring about a complete pacification of the
country, and at the same time win the friendship of the Awellimiden
chief, we ought to strengthen his hands.
 With this idea in my mind I make the following suggestions. We
should arm the Awellimiden with a hundred or a couple of hundreds
of percussion rifles, with very large nipples, which would only admit
of the use of special caps turned out in French manufactories.
With one hundred such guns the Awellimiden would be invincible,
and could soon butcher all their enemies, whether Kel Gheres or
Hoggars.
The absolute necessity of having French percussion caps would
place them entirely in our hands, and by doling out the ammunition
needed little by little, we should force them to submit to and serve
us. We should, moreover, have it in our power to break up their
strength directly they showed any reluctance to fall in with our
wishes.
In return for a service such as this supply of fire-arms, the
Amenokal would protect our traders; he has already in fact promised
to do so, not only by word of mouth but in writing.
These traders must, however, act with prudence and
circumspection. I am quite convinced that I and my companions
might fearlessly return to the Awellimiden because they know us
now. I have suggested to our Government that we should return, but
I have not been more successful in that direction than I have in
getting the rifles I asked for.
Strangers must not attempt with a light heart to penetrate into the
Tuareg districts, without having secured the formal protection of the
chief.
What would you have? When a Grand Duke announces his
intention of visiting the wine-shops of the outlying boulevards, don’t
we always take care to send an habitué of those boulevards with him
to look after him? A Jaume or a Rossignol[8] is always in attendance.
And if a protector is useful in Paris, can we not well understand that
one would be indispensable in the Sahara?
When Madidu has once said to a traveller “Yes, come,” or “You
can go,” I am convinced that no danger would be run in the districts
subject to him.
 With the Awellimiden on our side we could conquer the Sahara,
and the Tuaregs would help us to push on towards Lake Tchad, Air,
Tunis and Algeria. He would find it to his own advantage to do so,
and the conditions of his existence would be manifestly ameliorated.
Do you imagine that these Tuaregs are stupid enough to miss a
chance of getting stuffs for clothes, coverlids, glass beads, and all
the things they covet? If the men were sufficiently blind to their own
interests, I’ll warrant you their wives would not be.
The Tuareg race will be tamed at last, their faults, all the result of
the fierce struggle for existence, will disappear, and modern
civilization will have conquered a new district in Africa!
One afterthought does, however, occur to me. Will the change be
a good thing for the Tuaregs themselves?
When I think of their wandering life, free from all restraint, when I
remember their courage, which to them is the highest of virtues,
when I consider how truly equal all those worthy of equality are, I ask
myself whether after all they are not happier than we Europeans?
Their life is a hard one, and their habits are frugal, but has not
custom made this life natural to them, and are they really sensible of
its privations?
Good fortune with them is the reward of the brave who know how
to win the victory, and it is in razzis that the victory is gained. To spoil
the vanquished is also to wash out the stain of an hereditary injury,
for the vendetta is not confined to Italy, but often makes friendship
impossible between certain tribes in Africa. The goods of him who
perishes by the sword are the property of the wielder of that sword,
and the death of the vanquished avenges some pillaged or
massacred ancestor, as well as enriches the conqueror.
A rough rendering is given below of the Song of R’Otman, quoted
by Duveyrier, who justly calls it the Tuareg Marseillaise, which is
chanted in defiance of the Chambas by the Azgueurs, who are their
hereditary enemies.
 Death to thy mother! Ma’atalla the devil is in thee!
Call’st thou the Tuaregs traitors, the men of the plain?
Ha! but they know how to travel, to fight in the battle,
Sally at morn and return in the evening again!
Aye, and they know how to fall on the enemy sleeping—
Sleeping at ease in the tent with his flocks at his side,
Lapped in his fine woollen garments, his curtains and carpets
Spreading full length in the shade of the canopy wide.
What though with milk newly-drawn from the udders of camels,
What though with meat and with butter his paunch he has filled,
Straight as a nail to the ground pins the lance of the victor,
Out with a shriek and a yell flies the soul of the killed!
Sunk in despair lies the heart-broken wife of the victim,
Scattered and vanished their goods like as water o’erspilled!

Wild manners truly do these lines describe, but they also express
proud and heroic sentiments. What will the Tuaregs gain by their
transformation into civilized people?
In a few centuries, where the tents of the Amezzar are pitched
there will be permanent towns. The descendants of the Ihaggaren of
the present day will be citizens. There will be nothing about them to
remind their contemporaries of the wild knights of the desert.
No more will they go to war; no more will they lead razzis to
ravage the camps of their neighbours, for they will have given up
pillage altogether; but perhaps in a bank, which will take the place of
the tent of their Amenokal, they will try to float rotten companies, and
mines which exist nowhere but in the imagination of their chiefs.
What will they be then? Not pillagers but thieves!
Truth to tell, I think I prefer my marauders, who fall on their prey
like the lion Ahar!
AN AFRICAN CAMEL.
AN ISOLATED TREE AT FAFA.
 CHAPTER VI

FROM FAFA TO SAY

Our dread of the passage of the river at Fafa may have seemed
almost childish, and we have since had experience of many another
like it, but for a first attempt it must be admitted it was rather a
teaser.
Narrow and much encumbered, made more difficult by a violent
current, such is the pass of Fafa.
We took as guide the son of the chief of the village, who was later
to pay us a visit at Say. Thanks to him and with the help of his men
we crossed the first rapids without too much difficulty; but, alas! the
rope which was used to transmit to the rudder the movements of the
helm broke just as we emerged from them. Had this happened thirty
seconds sooner the Davoust could not have answered to her helm,
and would have been flung upon the rocks. The damage repaired,
we steered once more into the current, wending our way cautiously
amongst the numerous islands, skirting the course of the reef, our
good star bringing us safely into a quiet reach extending as far as
Wataguna, where we again came to flints lining the bed of the
stream.
In the evening we reached Karu, the Aube having struck once by
the way, but without sustaining much damage; still all these shocks
did not add to her waterproof qualities, and as she shipped more and
more water our anxiety and fatigue became greater and greater. We
had constantly to empty the hold, which did not conduce to the
repose of the passengers, who were often woke up by the noise we
made with our buckets.
 FAFA.

Karu is a pretty little village with thatched huts, amongst which

were many of the barns of a bee-hive shape used for storing millet
alluded to by Barth. We had noticed a good number during the last
few days. The inhabitants of this village are Rimaïbes or serfs of the
Fulahs and Bellates or slaves of the Tuaregs.
The chief of the latter told us how glad he was to see some white
men before he died. He added that he would like to give us some
sheep, but he understood that we never ate anything except the
flesh of black animals, and he had none of that colour.
I said that the colour of the wool did not trouble us at all; all we
cared for was the quality of the flesh, and he went and fetched us a
fine ram. It was the marabouts, who, to add to the probability of their
report that we were sorcerers, had made this assertion about black
animals. There is a custom in the Sudan that animals given as
presents should be as white as possible, as a sign of peace between
donor and receiver. We were now told that Bokar Wandieïdu, chief of
the Logomaten, had assembled a column of troops and was about to
attack us.
 At Karu the mountains were
pointed out to us which line the
famous rapid of Labezenga, which we
expected to reach the next day. A
guide was given to us who was said to
be wonderfully clever, but we saw no
particular sign of his intelligence.
It was on March 14 that we first
saw the terrible Labezenga rapid, and
KARU WITH MILLET I am very sure that we shall none of
GRANARIES. us ever forget it.
Our guide began the day by
performing a number of mummeries, the aim of which appears to
have been to make various evil genii propitious to us. From a leather
bag he took out a lot of flat and shaly flints which had been picked up
in the rapid. He wrapped each one of these flints in a separate piece
of cloth, spat upon them, and arranged them here and there all over
the boat.
The current rapidly swept us into a part of the river pretty free
from obstruction, and every now and then I tried to distract our
guide’s attention from his spells and to get him to give me a little
information, but he merely replied without looking at me that there
was no danger, and that he would stop us at the right time.
 THE LABEZENGA RAPIDS.

Often from behind some little jutting out point which intercepted
our view I heard a peculiar noise, a sort of dull but vague roar. The
rate of the current too increased rapidly, and we rushed along at a
rate of five miles an hour at the least. We listened eagerly, but all of a
sudden we saw that the stream was barred from side to side, a
distance of something like a thousand yards, by a positive wall of
rocks against which the water was dashing up in foam.
Our idiot of a guide looked up at last and saw the danger. He
motioned to us to steer for the bank, but rushing along as we were
with the tremendous current, to attempt to do so would have been
merely to drift helplessly on to the line of rocks, so we continued to
dash on with a speed which almost made me giddy, and presently, to
my intense relief, I saw a place on the right where there was less
foam. Yes, it was the pass, it was the gate of safety, we must make
for it, but was there any hope of our reaching it?
Our coolies bent to their oars and rowed so hard that they were in
danger of breaking them, whilst the sweat poured down their shining
black skins. I had just time to hoist the signal “Do as we do!” which
most fortunately Baudry and the captain of the Dantec understood.
They were just behind us. Now up with the oars and trust to our luck!
The speed increases yet more, the stream sweeps the boat towards
the pass, where it flings itself into the lower reach: we feel ourselves
falling, we shudder, we realize the fatal attraction drawing us in the
direction of the whirlpool; then like an arrow we shoot safely through
the opening. All is well with us at least. Our next anxiety is for our
comrades; we look behind, and a cry of terror bursts from our lips.
The Dantec, which is the next to attempt the pass, has stopped
suddenly; her mast is swept asunder, and has been flung across the
bow by the violence of the shock. All the men were thrown at the
same moment to the bottom of the boat, for the unlucky barge, which
had tried to pass about three feet on one side of the place where we
had got safely through, had struck against a rock which was hidden
by the whirling foam. She received a tremendous blow, but
fortunately did not sink.
But where was the Aube? That was our care now. She was
approaching rapidly, borne on by the current, but the whole pass was
blocked before her. She would crash into the Dantec, and both
vessels must inevitably be wrecked.
But no! Clouds of spray dash up over bow and stern alike; Baudry
has flung out the anchor and the grappling-iron: oh that they may
grip properly!
Thank God! They have. The Aube stops short some three
hundred yards at least from the Dantec at the brink of the rapid.
But what in the world is up now? The Aube is tilted at an angle of
some 45 degrees! The force of the current is such that it has taken
her in the rear and forced her into this extraordinary position, whilst
the grappling-chains and those of the anchor are strained to the
uttermost, producing the terrifying result described.
I now moored the Davoust to the bank, for we must try to save our
other boats.
 With regard to the Dantec it was a simple affair enough, for she is
a wonderful little craft, answering readily to the helm, and so buoyant
that we got off with no worse damage than the bursting asunder of a
couple of planks of her bottom. I sent Digui to help the men on board
of her, and she got safely through.
The rescue of the Aube was a more difficult matter, especially as
her rudder had got broken in the struggle. The anchor was raised all
right, but when it came to the grappling-iron we could not make it
budge; it had probably got jammed between two rocks, and all our
efforts to move it were in vain, indeed they only seemed to fix it more
firmly.
Driven on by the wind and whirled round by the strong eddies of
the current, the unfortunate barge began to describe semicircles
round her own grappling-iron. Of course when we once cut the chain
there would be no time to steer her, and we must therefore manage
to divide it exactly at the moment when she was opposite to the
opening she had to pass through. One second too soon or too late
and she would be lost.
I had climbed to the top of a little ridge, and with fast beating heart
I watched Baudry making his dispositions for the manœuvre he had
to attempt. A Tuareg chose this moment of awful suspense to tap me
on the shoulder and greet me with the formal salutation, Salam
radicum mahindia, and you can imagine how much notice I took of
him.
Without being at all put out by my silence, however, he went on—
“I see that you are in trouble. I have watched all that has been
going on from my camp behind the hills, and ever since early
morning I have felt sure that you were all lost. But God has saved
you and your people. I have forbidden my tribe to come and bother
you, for you know that we always beg of every one. Well, I am going
now, but if you have need of us, Tuaregs and negroes alike are
ready to help you, you have only to send me a messenger. Our
Amenokal has ordered us to meet your wishes.”
 As he finished his speech, I saw Digui deal a great blow to the
chain of the grappling-iron. The Aube fell into the rapid, but she
could not avoid the rock on which the Dantec had struck already.
She strikes, and the whole of her starboard side is completely
immersed. Is she staved in? No, her speed is such that she rushes
on as if nothing had happened. She is saved. A moment later she is
moored beside the Davoust.
“Not so much as a hole in her, Baudry!” I cried.
“No, I don’t think there is,” he replied, “but we had a narrow
escape.” We overhauled her, and there was not a leak anywhere. In
fact, Baudry declared that her planks were really more watertight
than ever.
Then my Tuareg, who had not gone away after all, but whom I
had completely forgotten, spoke to me again: “Enhi!” he said, which
means simply “look!” but his great wild black eyes shone with
pleasure from out of his veil as if some piece of good luck had
happened to himself.
Now are these Tuaregs brutes? are they men who can only be
swayed by interested motives? What nonsense to say they are!
Where did the interested motives come in here? Would it not have
been better for him if our boats had all been sucked down in the
rapids? We ourselves and all our goods would then have been his
lawful prey.
May Providence only grant that I never find any of my fellow-
countrymen worse than the Tuaregs.
You may be sure the brave fellow got his parcel of goods and
many other things as well. With his long swinging step he went off to
his people again, shouting to us by way of adieu, “Ikfak iallah el
Kheir” (“may God give thee all good things!”)
This was, however, but the first of the Labezenga rapids, and that
the easiest. We had scarcely gone a hundred yards further when we
came to a regular cataract some two feet high, barring our passage.
On one side rose lofty heights, on the left the stream was broken into
several arms by islands. In fact, there did not seem to be any
opening on either side, and we were all but in despair of getting
through this time.
Baudry spent the whole afternoon with our guide from Karu,
seeking a practicable pass, but everywhere the scene before him
was most forbidding, one cataract succeeding another and
alternating with boiling whirlpools, whilst the current rushed on at a
rate of seven or eight miles at the least. The river simply seems to
writhe in its course, and here and there it dashes backwards and
forwards from one side to the other of its bed as if in a state of
frenzy. There must be a difference of something like seven feet in the
height of the water.
The least impracticable place seemed to be on the left of our
anchorage between two islands, but I never should have believed
that any boat could pass through even that. We had, however, to
make the venture, and any delay would only render it more difficult,
for the water was falling rapidly.
On the morning of Sunday the 15th
Father Hacquart celebrated mass and
we then prepared for the passage.
The crew of our two big barges was
not strong enough to navigate both at
once, so we decided to send each
vessel separately past the dangerous
THE ‘AUBE’ IN THE RAPIDS. spots, supplementing one crew from
the other, and later we always
adopted this plan, which worked well on emergencies.
Digui was the only one of our captains who could manage such
tours de force, for really there is no other word for the work he had to
perform. Idris, the quarter-master of the Aube, rather loses his head
amongst the rapids, and is absolutely no good as a leader. Of course
all that can be done is to give a general indication of the course to be
pursued, and when the manœuvre has once begun everything must
be left to the intelligence of the pilot, and Digui alone of all my men
was really worthy to be trusted at the helm.
 We fortified ourselves with a good cup of coffee, feeling that it
might be our last, and the Davoust started, Baudry following us in a
canoe.
The scene before us was very much what it had been the day
before—a narrow pass, a diabolical current producing an impression
of unfathomable depth, which made our hearts sink and our breath
come in gasps. On either side the water whirled and surged and
roared unceasingly as it dashed over the huge rocks. Suddenly there
was a tremendous shock, and the boat seemed to slide away from
under our feet. It was the Davoust’s turn to-day. A hidden rock had
battered a hole in her bow in my cabin. Through the gap, some 20
inches big, the water came in in floods, and in less than ten seconds
it was a couple of feet deep.
But it was written in the book of fate that we were to go down to
the sea in the Davoust, and in spite of all our misfortunes, in spite of
everything being against us, in spite of reason, in spite of logic,
something always turned up to save us even at what seemed the
very last moment. The expected miracle always happened, and it is
no exaggeration to say that we experienced dozens of such
miracles.
We were going at such a rate when we struck the rock that for
one instant the barge remained as it were suspended on it, but the
next it was over it and in deep water again.
It so happened, as good luck would have it, that my servant
Mamé was in my cabin when the boat struck, and the water rushed
in at his very feet.
For the brave fellow to tear off his burnous, roll it into a ball and
shove it into the gap in the planks was the work of a few seconds;
that is to say, of just the time during which the rock held us fixed,
preventing us from settling down. We were saved once more. The
miracle had been performed. Only do not fail to notice what a
combination of circumstances was required to bring about the result:
the immense speed with which we were going making us actually
mount the rock, with the presence of Mamé in my cabin all ready to
stop up the hole!
 The Dantec passed through with us without difficulty, and it was
now the turn of the Aube. Digui attempted a manœuvre with her of
positively extraordinary audacity. Knowing all too well that the rock
which had been nearly fatal to us could not possibly be evaded, he
simply flung the boat upon the grass-covered bank, and she climbed
up, driven on by the great speed of the current. Then he let her slide
down again backwards, or, to use the strictly nautical term, to fall
astern.
For all this, however, we every one of us had to pay toll in one
way or another at this infernal Labezenga. The Aube grated on the
point of a hidden rock just as she was about to join us again in quiet
water.
It was now two o’clock in the afternoon, and we had been eight
hours getting over a little more than half a mile in a straight line. We
were famished with hunger, and our craving for food became almost
unbearable. I constituted myself cook, and drawing upon our
reserves of tinned meats and preserved vegetables, which we all felt
we were justified in doing under the circumstances, I seized what
came first, and tumbled everything helter-skelter into a saucepan.
We all devoured the result, which I called tripes à la Labezenga,
without in the least knowing what we were eating. I will give the
recipe to all who wish to emulate Vatel: tripes à la mode de Caen,
truffles, esculent boletus, haricots verts, with plenty of pepper and
spice, served hot. In N. Lat. 14° 57′ 30″, after just escaping from
drowning or from death in the jaws of a crocodile, nothing could be
more delicious, but somehow I have never ventured to try my olla
podrida again in France.
After a little rest, which was indeed well earned, Baudry went with
Digui to the village of Labezenga to try and get guides. He came
back in a state of terror at what he had seen.
For more than a month we had to lead a life such as I have just
described. What I have said will give an idea of all we went through. I
don’t want to dwell too much on our sufferings now that they are
over. Once embarked on such an enterprise as this there is nothing
for it but to go straight ahead, and by degrees one gets accustomed