Knowing by Perceiving Alan Millar

Visit to download the full and correct content document:

https://ebookmass.com/product/knowing-by-perceiving-alan-millar/
Knowing By Perceiving
Knowing By
Perceiving

Alan Millar

1
3
Great Clarendon Street, Oxford, OX2 6DP,
United Kingdom
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 in certain other countries
© Alan Millar 2019
The moral rights of the author have been asserted
First Edition published in 2019
Impression: 1
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 licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries 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
Published in the United States of America by Oxford University Press
198 Madison Avenue, New York, NY 10016, United States of America
British Library Cataloguing in Publication Data
Data available
Library of Congress Control Number: 2018952833
ISBN 978–0–19–875569–2
Printed and bound in Great Britain by
Clays Ltd, Elcograf S.p.A.
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
Contents

Preface vii

I. Epistemolo窃 and Philosophical Method

l. Our 知让liarity with knowledge
2. Pcrc叩tual knowledge 4
3. Why is ii 小at enquiry aims a心no心dge rather 归n meretrueKli叩 7
4. Scepticism about cone叩tua] enquiry 14
5. Phil伪ophy'$ inm心l in O心plS 19

2. Justified Belief, 氐asons., and Evidence 23

l. Justi.ficdbelief 23
2. `JUSl1fied,beliefin GeUler c豁es and well-fo皿ded belief 29
3. Knowledge bas叫 on evidence 33
4. Standing knowle吨e of fa巫 36
5. Cinching 的dence 38
6. Summary 42

3. Perception, E.'7perience, and Dir釭tR叫ism 43

1. Perceiving thin穿 directly 43
2. A non-rclationahst perspective SO
3. The relallonalist challtn穿 54
4. 压perienbnl orientation 56
5. Perception, co11SCiousness,at1d sensations 61
6. Non七ommi血l desrnptioos of 叩erience 66
4. Perceptual Kno叫 edge and Recognitional Abilities 73
l. Experiential a`｀屯reness and appear.u1ces 73
2. Percei寸ng, thinking, and recogni功ng 77
3. Pere叩lu吐-recognillonal abilities 83
4. Pere叩血l-r如gnilional abilities and the theory of knowledge 89
5. Perceiving-that and r叩心血lonal content 91

5. Perception and the Justification of BeUef 97

I. Perception and 叩ricnc:e 97
2. Do experiences justify by being evidence? 99
3. lrnmediate justification 106
4. Pcrcc:pl归1 k_nowlcd驴卫cll-found如css, and rational constraint I 09
5. Access to reasons constituted by trulllS as to 呻at we perceive Lobe so 112
6. Thc}U.s仙cation for 比lief I ha1 perception can yield 114
7. McDo-w'eIIon bo11ndaries 118
8. McDowell on intu.itional content 121
vi CONTENTS

6. Abilities, Competences, and Fallibility 125

J. Introduction 125
2. Some ordinary abilities 126
3. Havillg an ability and being reliably successful 129
4. Performances: hard and completable 131
5. Mental acts 136
6. Sosa and competence 143
7. Abilities: Further Issues 147
l. Cognitive contact and 山e reach of reason 147
2. Abilities in virtue epistemology 149
3. More basic abilities? 154
4. Retreating to dispositions plain and sitnple 157
5. Individuating recognitional abilities: some complexities 160
8. Knowledge from Perceived Indicators and Background Knowledge 165
I. Introduction 165
2. rndicators and evidei1ce 166
3. The covering generalization model 168
4. How indicators can provide clinching evidence 171
5. A puzzle about generalizations 174
6. The standit1g of accepted generalizations 176
7. More on back伊0皿d beliefs 181
8. Moore's truisms 184
9. Reason's work again 186
9. Going By What We Know 188
l. Rel伽g on what we know about knowledge 188
2. How can we gain so much from so little? 193
3. Self-understanding and the engaged standpoint 197
4. Not so radical scepticism 200
5. Radical scepticism and responsible judgement 205

References 209
Index 219
Preface

Perceptual knowledge, the central topic of this book, is a paradigm of knowledge.

When we know something perceptually we grasp a truth in a rational intellectual act.
That act is the exercise of a general ability to come to know truths of a certain sort
perceptually. If I see that, and in that way know that, the person entering the room is
my brother, then I exercise a general ability to recognize my brother as my brother
from the way he looks. This ability is nothing less than a knowledge-acquisition
ability. It consists in being able to tell, that is, come to know, concerning my brother
that he is my brother from the way he looks. Since abilities are exercised only in doing
what the ability is an ability to do, it is exercised only when I acquire such knowledge.
Knowledge gained from the exercise of abilities such as this is readily available to us
and its contents extend well beyond superficial features of things such as colour and
shape. I recognize my brother as my brother and not merely as a thing with a certain
visual appearance.
Epistemological discussions of perception usually focus on something other than
knowledge. They begin with consideration of how beliefs arising from perception can
be justified. With the retreat from knowledge to justified belief there goes a retreat
from perception to the sensory experiences implicated by perception. Indeed, on the
most widely held approach to the topic, perception drops out of the picture other
than as the means by which we are furnished with the experiences that are supposed
to be the real source of justification—experiences that are no different in kind from
those we could have had if we had been perfectly hallucinating. My first book,
Reasons and Experience (Millar 1991a), was very much in this spirit. I sought to
challenge Donald Davidson’s (1983/2001) claim that experiences (which he called
sensations) could not justify beliefs. This general approach to epistemological issues
concerning perception now seems to me to be mistaken.
The motivation towards accepting a different view derived from consideration of a
conception of perceptual knowledge, and a related conception of how to do epis-
temology, both of which are at odds with much of recent epistemology, but seemed to
me to be eminently plausible. A significant element in this alternative way of thinking
is a conception of reasons and justified belief. On this conception, reasons to believe
are constituted by propositions or, more precisely, by truths. A reason to believe that
P that is constituted by the truth that Q can be your reason for believing that P only if
you believe that Q and you believe that P for the reason that Q. On this way of
thinking we can make good sense of what it takes for one to be justified in believing
something: you are justified in believing that P only if there is an adequate reason for
you to believe that P and you believe that P for that reason. Under this conception—
call it the reasons conception of justified belief—we are not at a loss to understand how
viii 

it can be that some things we believe can constitute an adequate reason to believe
something else. For they might be such that we can validly deduce something else
from them or such that they present strong evidential support for something else.
Once we move away from the idea that justified belief has to do with propositionally
constituted reasons it is far less clear how what is supposed to provide justification
can do so. In particular, it is less clear how an experience or course of experience can
provide justification for a belief.
This was an issue I addressed in Reasons and Experience. I exploited the idea that
having a justified belief that P is a matter of competently believing that P. Competence
was to be understood as competence in employing concepts. Just as believing some-
thing on the basis of other things one believed could be an exercise of conceptual
competence, so believing something on the basis of an experience could be an exercise
of conceptual competence. I came to reject this view, largely because it is obscure how
experiences can stand to things believed in any relation that is truly analogous to the
way in which a (propositionally constituted) reason to believe something can stand to
that thing. The very idea of competently basing a belief that P on the assumption that Q
presupposes that the assumption that Q stands in an adequate reason-giving relation to
the proposition that P in that, if it were true, it would constitute an adequate reason to
believe that P. Invoking competent basing would not explain justification, unless it
were itself explained in terms of the exploitation of a grasp of reason-giving links. It is
obscure how experiences can stand to things believed in anything like the reason-giving
relation in which assumptions can stand to things believed.
It seemed right to hold onto the reasons conception of justified belief if it could be
shown to serve our epistemological goals. It is, after all, tolerably clear and arguably
in keeping with common-sense thinking. With respect to a wide range of our beliefs
we have a working idea of how to judge whether they are held for adequate reasons.
Also significant was the realization that the reasons conception is compatible with an
eminently plausible account of perceptual knowledge. This account, on which
I touched above, accords a central role to general abilities to come to know truths
of certain sorts. More specifically, it accounts for perceptual knowledge in terms of
the exercise of perceptual–recognitional abilities. Those abilities are in the neigh-
bourhood of what I took to be conceptual competences in that they have to do with
the mastery of concepts, but they are nothing less than abilities to tell, that is, come to
know, of things that are some way that they are that way from the way they appear to
one or other of our sense-modalities. In this respect they differ not just from
competences as I previously conceived them, but also from the competences or
abilities that have been invoked by contemporary virtue epistemologists, notably
John Greco and Ernest Sosa. In taking this line I unashamedly abandoned the idea
that perceptual knowledge is to be explicated in terms of an independently graspable
conception of perceptually justified belief. Perception does give rise to justified belief
but, to explain how, we need an account that differs from that which makes
justification by experience central. (See Chapter 5.)
  

I was struck by, and drew encouragement from, Timothy Williamson’s (2000)
reversal of the usual way to think of the relation between knowledge and justification,
and his scepticism about the prospects of giving plausible reductive accounts of
knowledge in general. It came to seem obvious that, at least in its central sense,
justification for belief is conferred by knowledge. This meshed with the idea that in its
central sense a belief can be justified only if it has a sound foundation. How could a
sound foundation be provided by something less than what we know? Yet this way of
thinking raised a question as to whether anything of interest can be said about
knowledge in general. I suspect that much of the resistance to Williamson’s view—
the so-called knowledge-first view—derives from thinking that it stands in the way of
saying anything substantive about what knowledge is. Williamson himself interest-
ingly explores how knowledge links up, not just with justified belief, but with, for
instance, assertion, evidence, safety, and reliability, but one reason why resistance
persists, aside from the friction provided by tradition of several decades, is that many
epistemologists are dissatisfied with anything less than a substantive general account
of knowledge, and preferably one that is reductive.
Rather than attempting to formulate a highly general account of knowledge and
then attempting to shed light on knowledge of the various kinds, we might do better
by investigating the kinds with a view to illuminating what knowledge is in general.
Perceptual knowledge is a kind of knowledge because it is characterized by a
distinctive way in which the knower grasps the truth or fact that is known. The
same applies to a certain kind of knowledge from a perceived indicator (sign), as
when one knows that fire is near a location because one recognizes that the presence
of smoke is a sign that fire is near that location. I argue (Chapter 8) that when one has
recognitional knowledge of what is indicated by a perceived indicator, the knowledge
gained, though evidence-based, is more like perceptual knowledge than one might
initially think.
Of course, if we are able to tell of knowledge of some kind that it is knowledge we
must have some general conception of knowledge. Perhaps the best we can do to
articulate that conception is to think of knowledge as being in cognitive contact with
a truth or fact. We may seek to illuminate what that cognitive contact amounts to by
giving substantive accounts of the diverse forms that it can take. My focus on
perceptual knowledge is in keeping with this conception of how to proceed.
The bewildering variety of intuitions, theories, and methodological predilections
in contemporary philosophy demands that in making a case for the conception of
perceptual knowledge developed here, one has to consider many other issues,
including issues about method in philosophy, and in epistemology in particular.
Much of the discussion in this book is, accordingly, not just about perceptual
knowledge but about topics giving rise to positions that can be obstacles to finding
plausible what I have to say about perceptual knowledge.
My thinking about perceptual knowledge has been shaped by engagement with the
thinking of disjunctivists (Millar 2007, 2008a, 2008b). At the heart of disjunctivism is
x 

a relationalist conception of perceptual experiences—the sensory experiences that are

implicated in episodes of perceiving mind-independent objects. The core idea is that
any such experience consists in awareness of (acquaintance with) mind-independent
objects. Disjunctivists acknowledge that there can be sensory experiences that are
not such episodes, but which are such that it is just as if one is perceiving mind-
independent objects. For them, experiences of very different kinds can fall under the
description ‘being such that it is just as if one is perceiving an F’. Indeed, disjuncti-
vism gets its name from the claim that an experience such that it is just as if one is
perceiving an F can be either an episode of awareness of an F or some other kind of
episode, for instance, a visual hallucination. I have never found this relationalist
conception plausible. Nonetheless, the thinking of disjunctivists about epistemo-
logical issues, notably in the work of John McDowell (1982, 1994, 1995), seemed to
me to give rise to serious challenges to familiar ways of thinking about how
perception yields knowledge and justified belief. If the contribution that perception
provides to one’s knowing that a spider is before one is a visual experience that one
could have had when one was not perceiving any spider but merely hallucinating,
how could the fact that a spider is before one be visually manifest to one—a fact that
one takes in, knows, perceptually? When one knows, because one sees, that a spider is
before one it is integral to this knowledge, and indeed to the justification one has for
believing that a spider is before one, that one is in perceptual contact with the spider
and in cognitive contact with the fact that this particular spider is there before one.
This exemplifies what I take to be the profound truth in disjunctivist thinking so
far as it concerns perceptual–demonstrative knowledge about currently perceived
objects. There is, though, a question as to whether this truth makes sense only on the
assumption of the relationalist conception of perceptual experience. I make a case for
thinking that it is not inextricable from that conception (Chapters 3–5 of this book
and Millar 2018). The case turns crucially on the distinction between perception and
the experiences implicated by perception. Relational episodes fill the key explanatory
roles, but these are ordinary episodes of seeing, hearing, and the like. To defend this
view we need to show that it is intelligible that such episodes should fill this role even
if the relationalist conception of experience is false.
I said that perceptual knowledge is a paradigm of knowledge. What makes it a
paradigm is that it consists in a grasp of a truth in an act of perceptual recognition.
There is a tight relation here between something we do and a fact—we exercise a
recognitional ability and in doing so come to know that fact. If this is exemplary of
how we relate to facts we know, one might reasonably wonder whether much that we
tend to think of as empirical knowledge can pass muster as knowledge. Reflecting on
beliefs one holds on the basis of empirical evidence, by contrast with beliefs that one
holds simply in virtue of perceiving that something is so, might well lead one to
wonder whether evidence ever truly settles or establishes that something is so, as
opposed to justifying, at best, a high degree of confidence that it is so. In response,
one might adopt an all-out Russellian Retreat (Wright 1991) and focus on more or
  

less justified belief rather than knowledge, or one might take the realm of knowledge
proper to be very severely restricted (Bonjour 2010).
Addressing similar issues, Charles Travis (2005) invoked a notion of factive
meaning. According to Travis it could be that a waft of perfume about Sid factively
means that he is seeing someone. If the waft of perfume does mean this, then Sid is
seeing someone and if Zoë is suitably au fait she could come to know this on smelling
the perfume. Though I was critical of how Travis developed this idea (Millar 2005),
I came to think that he was onto something important. A state of affairs or event can,
as I put it, indicate (be a sign) in a factive sense that something else is so, as in the case
in which the occurrence of smoke at a place indicates that fire is nearby. It is not just
that the presence of smoke is a reliable indicator of the nearby presence of fire. The
point is that since there is smoke there is also fire and, further, that in the circum-
stances there would be this smoke only if there were fire. A person who has learned to
recognize that smoke indicates fire could have more than merely some support for
thinking that fire is nearby on seeing smoke. Such a person can have clinching
evidence of the presence of fire provided by the presence of smoke.
Once one begins to take seriously the idea that evidence adequate for knowledge is
clinching evidence a host of problems emerges. Some of them concern the very idea
of clinching evidence and how it can be exploited as such evidence. Others concern
how to make sense of our having the vast amount of factual knowledge that we have
acquired second-hand, as when we glean information from encyclopaedias or other
sources of information. I make tentative forays into this territory (Chapter 8) though
it calls for much more investigation, not only for reasons internal to academic
philosophy, but because we live at a time in which there are forces at work that
would undermine the means by which knowledge can be socially transmitted.
Chapters 1 and 2 of the book deal with method in philosophy and with key concepts
of evidence and justified belief. They set the scene for Chapters 3–5, which provide the
detailed treatment of perceptual knowledge and of the justification for belief that
perception can yield. The proposed account of perceptual–recognitional abilities
makes certain assumptions about abilities and their exercise. Chapter 6 is designed
to show that those assumptions are in keeping with a plausible treatment of abilities
in general. Some further complexities and puzzles concerning abilities are explored in
Chapter 7. Chapter 8 focuses on the kind of knowledge from perceived indicators to
which I alluded above. This topic naturally leads into consideration of background
beliefs and knowledge, and from there to discussion of the issue, on which I touched in
the previous paragraph, about factual knowledge gleaned from sources of informa-
tion. Chapter 9 has further discussion about method, focusing on the importance in
epistemology of relying on what we know about knowledge. It also explores in general
terms why knowledge is puzzling. It is, in effect, a plea not to be unduly swayed by
scepticism or allow the issue of scepticism to dominate our thinking about knowledge
and justified belief.

* *
xii 

Over the several years in which I have been working on topics that crop up in this
book, I have benefitted enormously from the philosophical companionship of col-
leagues at Stirling. I am grateful to all who have raised issues for me to consider and
for helping me towards clarifications of my views. Special thanks are due to Adrian
Haddock, with whom I have had numerous conversations and written exchanges.
The attention I have paid to first-person reflection at some points is in large measure
due to Adrian. Philip Ebert helpfully commented on the last chapter, and prompted
improvements. Peter Sullivan has an uncanny habit of getting to the heart of
philosophical issues. I am constantly grateful for his acute critical questioning and
philosophical insight. At various points Zoe Drayson, Antony Duff, Rowan Cruft,
Simon Hope, Kent Hurtig, Colin Johnston, Sandra Marshall, Peter Milne, Tony
Pitson, Sonia Roca, Alex Stathopoulos, and Michael Wheeler have raised questions
that called for further reflection on ideas I have presented.
Collaborating with Adrian Haddock and Duncan Pritchard, between 2005 and
2008, on the AHRC-funded Value of Knowledge project contributed a great deal to
the shaping of my thinking about perceptual knowledge and knowledge in general.
Ernest Sosa’s support for that project, his work in virtue epistemology, and his
interest in my own thinking, have been both encouraging and stimulating. I am
grateful, too, for helpful discussion over several years with philosophical audiences at
Aberdeen, Bled, Bologna, Cambridge, Copenhagen, Edinburgh, Fribourg, Geneva,
Glasgow, Helsinki, Hull, Leuven, Oslo, Paris, Rijeka, St Andrews, and Warwick.
I thank an anonymous reader for Oxford University Press, who provided comments
that led to improvements in the text, Peter Momtchiloff for his encouragement
throughout, and Jennifer Hinchliffe for her excellent copy-editing. Thanks are also
due to Malcolm Millar, who made helpful comments on a draft of this preface that
prompted revisions. Finally, I thank Rose-Mary again for her patience and support,
and for allowing me to use a detail from one of her paintings for the book’s cover.
Though most of this book comprises freshly written text, I have drawn on previous
publications in some chapters. I thank Oxford University Press and The Aristotelian
Society for permission to reproduce some passages from articles published in Pro-
ceedings of the Aristotelian Society, in particular, Millar (2011b and 2014b), which
appeared in, respectively, Supplementary Volumes 85 and 88, and Millar (2017), a
discussion note published in the ordinary Proceedings, Volume 117. The passages in
the Supplementary Volumes occur in Chapters 1 and 2. Those from the discussion
note appear in Chapter 7; this material is reproduced by permission of Oxford
University Press and reprinted courtesy of the Editor of the Aristotelian Society:
© 2011, 2014, and 2017 respectively. Chapter 6 reproduces with amendments and
additions most of Millar (2016a), which is my contribution to Performance Epistem-
ology: Foundations and Applications, edited by Miguel Ángel Fernández Vargas,
© 2016, reproduced by permission of Oxford University Press. I thank Oxford
University Press for their permission to use this material and to use in Chapter 8
some passages from Millar (2014a), my contribution to Scepticism and Perceptual
  

Justification edited by Dylan Dodd and Elia Zardini, © 2014, reproduced by permis-
sion. Finally, I thank Cambridge University Press for allowing me to reproduce in
Chapter 5 some passages from Millar (2016b), my article in Episteme, Volume 13,
Issue 1, © 2016, journals.cambridge.org/epi, reprinted with permission.
Bridge of Allan
September 2018
 1
Epistemology and Philosophical
Method

1. Our familiarity with knowledge

Propositional knowledge—knowledge that such-and-such—is not an elusive com-
modity. We routinely seek to know, and often find out, that is, come to know,
whether some specified thing is so. We often know that we have succeeded in finding
out what we sought to know. The scope of our knowledge might well be problematic
but it is not seriously in dispute that we know a great deal that relates to the
practicalities and curiosities of everyday life. We can know when a train is scheduled
to arrive or depart by reading the official timetable. (Knowing whether the train will
arrive or depart at that time is a different matter.) We can know that friends are on
their way to meet us because they have phoned to tell us. We can know by looking in
the fridge that we are out of eggs. We can know that it is safe to touch a bare electrical
wire because we know we have just switched off the power at the mains box. This last
example serves to make vivid that we often settle for nothing less than knowledge. If
we are to touch the wire we want to know that the power is off, not just to have some
reason to think that it is likely to be off. We could know this by looking at the mains
box ourselves or from someone else’s telling us that that he or she has just switched
off the power.
A vast amount of what we know we have found out by using simple methods, for
instance, by personally looking or listening or tasting, or by being informed by others.
Sometimes we tell that something is so by recognizing that something we observe
indicates that it is so. (See Chapter 2, sec. 5 and Chapter 8.) For instance, one might
tell, so find out, that it has been raining from the water dripping from trees and
bushes and the surface water on roads and pavements.
In trying to shed light on what knowledge is we have to engage with knowledge as
we find it, for how else are we to construct models that accurately represent what
knowledge is? Here, as elsewhere in philosophy, we need to exercise our imaginations
in thinking about scenarios illustrative of our topic and how we might react to those
scenarios. In doing so we must select realistic examples that reflect the kinds of
knowledge we know we have. We need to draw upon relevant experience of instances
of knowledge-acquisition and knowledge-transmission. Experience such as this

Knowing By Perceiving. Alan Millar, Oxford University Press (2019). © Alan Millar.
DOI: 10.1093/oso/9780198755692.003.0001
    

supplies us with the materials for our imaginations to work with. Juggling with
theories, and ingeniously defending positions, is of no avail if we lose touch with
the realities. (For further discussion on this theme, see Chapter 9, sec. 1.)
The aim, then, is to consider what knowledge must be in the light of truths of a
familiar sort pertaining to knowledge. In pursuing this I recommend that we first
enquire into particular kinds of knowledge, for instance, perceptual knowledge or
knowledge from testimony, and try to understand knowledge in general in the light
of these more specific enquiries.¹ (On kinds of knowledge, see Chapter 4, sec. 4.) In
mainstream epistemology it has been assumed that the various kinds of knowledge
are best understood in the light of an analysis of knowledge in general. Indeed, much
of the work in this tradition focuses on highly general issues at the expense of detailed
treatment of particular kinds of knowledge. The Core Thesis is that knowledge is to
be understood in terms of conditions on belief. The analysis sought would be
complete in that the conditions would be necessary and sufficient for knowing. It
would be reductive in that grasp of the analysans would not require one to have an
independent grasp of the concept of knowledge. Prominent among the conditions
proposed is that the implicated belief should be justified. Justification, though, is very
diversely conceived—in terms of reasons, in terms of evidence, in terms of some
broader notion of right or entitled or rational believing, and even in terms of being
formed by a mode of belief-formation that would reliably yield true beliefs. (See
Chapter 2, sec. 1.) A further complication is that, in keeping with the Core Thesis, the
justification condition might be supplemented or replaced by something else. An
important idea is that a belief of a certain sort (with a certain content) counts as
knowledge only if formed in a manner that would reliably yield true beliefs of that
sort (with that content).
A. J. Ayer in The Problem of Knowledge (1956: 8–10) was explicit that the analysis
sought would be concerned with the meaning of the term ‘knowledge’. Since Ayer’s
preoccupation was not with a particular natural language, this way of thinking
focused in effect on the concept of knowledge—a concept that is expressed by the
English term but also by corresponding terms in other languages. It was by reflection
on this concept that we were to arrive at conditions under which a person would
know this or that. The quest for conceptual analysis has its roots in the logical
positivist preoccupation with linguistic meaning and with the demarcation of philo-
sophical enquiry from empirical enquiry.² The primary concern of philosophy would
be with analytic or conceptual truths rather than synthetic or factual truths. An
analysis of knowledge, for instance, would have the form, ‘X knows that P if and only
if X believes that P, and . . . ’. The upshot would be an analytic or conceptual truth—a

¹ For a similar attitude applied to the more specific issues, see Roessler (2009 and 2011).
² The preoccupation with meaning and with the demarcation of philosophy from empirical enquiry is
explicit in Ayer’s Language, Truth and Logic (1946, but originally published in 1936). See, especially, chs. II
and III.
     

proposition the truth of which is guaranteed by its content. Denying such a truth
would indicate failure either adequately to grasp the concept of knowledge or to
follow through the ramifications of its application.
That the analysis of knowledge should be complete reflects the idea that enquiry
into the concept of knowledge is enquiry into what knowledge is. The thought here is
that since knowledge is what falls under the concept of knowledge we can shed light
on what knowledge is by reflection on its concept. An analysis that fell short of
providing necessary and sufficient conditions would seem to fall short of a full
account of what knowledge is. The requirement of reduction reflects the idea that
the analysis should be explanatory in that it should explain what knowledge is in
terms we could understand even if we lacked a prior grasp of the concept of
knowledge. It was thought that if the analysans were articulated by using a term
for knowledge, this condition would not be met.
Despite fairly widespread scepticism about conceptual truth and the associated
conception of conceptual analysis, the idea that philosophy aims, among other
things, at complete reductive analyses has not been abandoned. Prudent philo-
sophers might be reluctant to present their analyses as conceptual. Some might
prefer to think of them, by way of contrast, as metaphysical. Others might not be
explicit as to what they are about. In epistemology, though, the project of complete
reductive analysis, however viewed, has had to reckon with the long, unpromising
history of tackling Gettier problems.
Edmond Gettier’s (1963) imaginary examples were designed to show that know-
ledge is not adequately explicated as justified true belief. By inspection of the
examples we were supposed to think it evident (a) that the subject in them has a
belief that is both true and justified, and (b) that this subject does not know the truth
that constitutes the content of that belief. Additional or alternative conditions were
proposed but were found to be unsatisfactory because new kinds of what came to be
known as Gettier cases were presented with a view to showing that the conditions
proposed were either not sufficient or not necessary. Despite the efforts over several
decades, it is fair to say that there is still no widely accepted analysis. As Timothy
Williamson (2000: 30, 32) noted, the history of analyses and counterexamples does
not encourage one to suppose that a plausible analysis of knowledge in terms of
conditions on belief is in prospect, and there is a serious question as to whether one
should expect there to be such an analysis.³ Williamson’s pessimism has not been
generally shared. An important factor, I take it, has been the seemingly irresistible
conviction that knowledge cannot be understood other than as belief that satisfies
certain conditions and the further assumption that it will not be adequately under-
stood except in terms of a complete reductive analysis. I suspect, too, that there has
been a degree of bafflement at the idea that there could be an illuminating and

³ For discussion, see Cassam (2009) and Williamson’s (2009) response.

    

satisfying account of knowledge that did not meet these desiderata. The discussion in
this book is intended to counter that bafflement. In this I am inspired by how
naturally it comes to us, in the course of ordinary (non-philosophical) life, to think
and speak of knowledge in terms that in no obvious way require to be grounded in, or
explained in terms of, more basic considerations about beliefs. Nowhere is this more
evident than in connection with perceptual knowledge.

2. Perceptual knowledge
In the sense in play here, perceptual knowledge is knowledge gained from what you
perceive by means of one or other of the senses of sight, touch, hearing, smell, and
taste. For each of these senses there is a way or mode of knowing.⁴ Visually
recognizing a tree as a birch, thus seeing that it is a birch, is a way of knowing it to
be a birch. It is constitutive of such visual–recognitional knowledge that it is acquired
in exercising a visual–recognitional ability. To have such an ability is to have a way of
telling, and so coming to know, of birches that they are birches, going by the way they
look. Similarly, you might be able to recognize a person’s voice as the voice of some
specified person from the way it sounds. To have such an ability is to be able to tell,
and so come to know, of that person’s voice that it is that person’s voice from the way
it sounds. In subsequent chapters I shall have much to say about abilities such as
these. I shall, among other things, consider what it is to exercise them, how their
exercise depends on there being environments favourable to their exercise, and what
it is for us to be fallible in relation to them. The point I want to emphasize at this stage
is how naturally it comes to us to think of perceptual knowledge in the terms I have
just described. We would not miss something crucial about someone’s knowing that
that tree is a birch from the way it looks for want of a conception of what justifies the
person’s belief that the tree in question is a birch, or a conception of the operative
processes leading to the formation of that belief. Yet, we are not powerless to
understand what it is to be able to recognize birches as birches from the way they
look or, as we might put it, what it is to know a birch when one sees one. Nor are we
powerless to understand how a person might have acquired this ability. He might, for
instance, have undergone training in tree management that involved learning to
identify various species of tree. The way we most readily think of knowledge gained
through the exercise of such an ability is in terms of the success concepts of knowing
and seeing, rather than in terms of belief and some notion of visual experience that
falls short of seeing. (On the latter, see this section below and Chapter 3, sec. 2.)

⁴ A notion of ways of knowing is central to the discussion in Cassam (2007). For discussion of this in
relation to Williamson (2000), see French (2014). Some ways of knowing, for instance, seeing that such-
and-such, can be illuminated only in terms of a way of telling, conceived as a means of coming to know. See
Chapter 4, sec. 4.
   

Perceptual knowledge is just one kind of knowledge, but it is suggestive of a more

general conception of knowledge: knowing is having a grip on reality, being in
cognitive contact with a fact.⁵ This general conception is to be understood in terms
of specific ways or modes of knowing. These in turn can be more or less specifically
characterized. Visual–recognitional knowledge is knowledge that something is some
way from the way it looks. Tactual–recognitional knowledge is knowledge that
something is some way from the way it feels.
There is a close relation between recognition and ways of telling. In acquiring
visual–recognitional knowledge you deploy a way of telling of things that are some
way that they are that way from the way they look. Since a way of telling just is a
way of acquiring knowledge of a certain sort, thinking of knowledge in these terms
will not serve the purpose of a reductive analysis of knowledge, but part of what is
at issue is whether we should assume that knowledge admits of such analysis. On
the project that I am recommending, the substantive work goes into, among other
things, making explicit the ways of knowing that individuate kinds of knowledge.
Alongside perceptual knowledge there is, for instance, knowledge from testimony
and knowledge from perceived indicators, as when you know that the postman has
passed by in view of the letters in your mailbox or on your doormat. The interest-
ing detail is to be found by close attention to those ways of telling, but there is no
commitment here to the view that all of our knowledge is acquired by our
deploying a way of telling. We soak up a vast amount of knowledge, not by
deploying ways of telling, but through being immersed in a community in which
this knowledge reposes, and being educated and acculturated into that community.
Some of the soaking up takes place in advance of acquiring the discernment that is
required for having command of ways of telling. This is a theme I pursue in
Chapter 8, sec. 6.⁶
That we readily employ the concept of knowledge is confirmed by consideration of
enquiry. The aim of enquiry into whether something is so is to find out the truth of
the matter. Finding out is nothing less than coming to know. This does not, of course,
establish that the finding out does not admit of reductive analysis, but rather than
presuppose that it does we should consider what if anything compels us to think in
this way. Perceptual knowledge is again instructive.
Suppose that I find my keys, so come to know where they are. What compels or
even invites me to think of what I have done in terms that implicate justified belief?
Could it be something about making sense of my having arrived at this knowledge?
Surely, it might be suggested, I think of my acquisition of the knowledge as a matter

⁵ The metaphor of cognitive contact is used in Putnam (1994: 454) and Zagzebski (1999: 109). A similar
metaphor, ‘cognitive purchase’, is used in McDowell (1995: 883). For discussion, see Millar (2007: secs. 4
and 5) and this volume Chapter 4, sec. 4, Chapter 7, sec. 1, and Chapter 8, secs. 1 and 2.
⁶ I discuss it in connection with detached standing knowledge in my contribution to Pritchard, Millar,
and Haddock (2010).
    

of having formed a belief on a certain basis. If I discovered my keys through looking

for them, and spotting them on the telephone table, then is it not natural to think of
myself as having formed a belief as to their location on the basis of a visual
experience? If visually experiencing something were just visually perceiving it there
would not be much wrong with this. Perceiving something is certainly a way of
coming to know about it, and therefore coming to have beliefs about it. The problem
is that the very idea of experience in the context of philosophical discussions of
perception is contested. A widely held view is that on perceiving my keys I have a
certain sort of experience—an experience such that it is just as if I am looking at my
keys. Crucially, though, having such an experience is conceived to fall short of
seeing my keys. To perceive the keys further conditions must be satisfied, including
the condition that the presence of the keys causes me to have the experience. As
I shall argue later, it is not obvious that experiences so conceived justify or
contribute to the justification of beliefs. (See Chapter 2, sec. 3 and Chapter 5.) In
the situation under consideration I certainly found my keys through looking for
them and then seeing where they were. This is not in dispute. The issue is whether
we need to think of the discovery in terms of forming a true belief on the basis of a
visual experience under the conception just described. Note that, whereas it might
be perfectly obvious to me that I found my keys and how I found them—I looked
for them and then saw them on the telephone table, recognizing them to be my
keys—it need not be at all obvious to me whether I count as having found them on
account of having formed the belief on the basis of a course of visual experience.
After all, I might not have been introduced to the distinction between seeing my
keys and having a visual experience such that it is just as if I am looking at my keys.
The very idea of a visual experience in that sense is relatively sophisticated. (See
Chapter 3.) It need not enter my thinking even though I well understand that, and
how, I found the keys. This matters because we should be able to make sense of our
ordinary evaluations of what people think. It would be difficult to make sense of
such evaluations if their correctness turned on factors of which we have no
conception. Yet, we surely can make sense of them. (See Chapter 5.) Thankfully,
it is readily intelligible that I should be thought to have found out where my keys
are and surely correct to think of me as having done so in virtue of having seen
them and recognized them to be my keys.
Nothing I have been claiming is at odds with the thesis that knowing that P entails
believing that P. It is plausible that in the order of understanding, knowledge is prior
to belief and that, as Williamson suggests, to believe that P is to have an attitude to
the proposition that P ‘similar to the ways in which subjects treat propositions which
they know’ (2000: 46–7). There is good reason to read this in the light of a liberal view
of how similar a case must be to knowing that P to count as believing that P. It would
be too strong to require that for subjects who believe that P it is just as if they know
that P. Failing to meet that requirement they might still use the assumption that P as
        ? 

a premise in reasoning, be prepared sincerely to assert it while not giving it to be

understood that they know it, and the like.⁷
Nor is anything I have claimed or assumed at odds with the idea that when one
acquires perceptual knowledge that P one manifests a belief-forming disposition such
that (a) by and large, beliefs formed in manifestations of that disposition are true and
(b) some beliefs formed in manifestations of that disposition are false. For all I have
said, it might be that when coming to know of a perceived F that it is an F from the
way it looks, the resultant belief concerning that F to the effect that it is an F is formed
in manifesting just such a disposition. There is a question, though, as to whether we
need to advert to belief-forming dispositions to account for how one comes to know
by means of visual perception, for instance, that something we perceive is an F. The
natural way to explain the acquisition of knowledge in such a case is in terms of the
exercise of a perceptual–recognitional ability—an ability to recognize Fs as Fs from
the way they look. Because of their fallibility, those who have the ability are likely to
form false beliefs occasionally. We have an idea of the sort of thing that would explain
false belief on the part of such a person, but we need have no clear conception of the
complex belief-forming dispositions that the person has. That is no obstacle to
understanding how those who have the ability to recognize Fs as Fs from the way
they look acquire the kind of knowledge in question. They do so by exercising the
ability. (For more on this theme, see Chapter 6 and Chapter 7, sec. 4.)

3. Why is it that enquiry aims at knowledge

rather than mere true belief?
A source of resistance to understanding any kind of knowledge in terms of ways of
telling derives from the deeply rooted idea that enquiry aims at true belief and that
this aim can be understood independently of consideration of knowledge. Alvin
Goldman clearly articulates the idea in the following passage.
. . . people’s dominant epistemic goal, I think, is to obtain true belief, plain and simple. They
want to be informed (have true belief) rather than misinformed or uninformed. The usual route
to true belief, of course, is to obtain some kind of evidence that points to the true proposition
and away from rivals. But the rationale for getting such evidence is to get true belief.
(Goldman 1999: 24)

Working with the assumption that enquiry aims at true belief, the tradition presses us
to consider why in enquiry we should want knowledge at all, rather than mere true

⁷ I do not think that asserting that P entails giving it to be understood that one knows that P. When we
tell others something we give it to be understood that we know, but asserting should not be equated with
telling. People make all sorts of pronouncements about politics, for instance, though they would not be,
and would not think of themselves as, telling, in the sense of conveying knowledge. Yet their pronounce-
ments might count as assertions and as expressions of what they believe.
    

belief, and to explain what is so good about knowing. One might suppose that the
answer is simply that going for knowledge, when knowledge is in prospect, gives us
our best chance of getting true belief. I think that this is the wrong way to view the
matter.
In Meno, Plato represents Socrates as saying:
. . . true opinions are also a thing of beauty, as long as they stay with one, and all their
consequences are good. But they’re not prepared to stay with one for long. Instead they run
away from the person’s soul. As a result, they are not worth very much until someone ties them
down by reasoning out the cause. . . . When they’ve been tied down, they become, first of all,
instances of knowledge, and, secondly, settled. It’s precisely for this reason that knowledge is
something more precious than correct opinion, and it’s being tied down that makes knowledge
different from correct opinion. (Meno 97e–98a)⁸

For present purposes it does not matter how Plato conceived of reasoning out the
cause. My concern is with the suggestion that stability of true belief has a role to play
in the explanation of the value of knowledge rather than with how Plato thought that
stability is achieved.
It is one thing to suppose that knowledge has agent-neutral value in that, whatever
its content and however one relates to its content, it is a good thing to have, and
another to suppose that knowledge on some matter has agent-relative value on
account of how the matter relates to some individual. It might well be that Plato
thought knowledge has agent-neutral value, but I think his suggestion works best in
relation to the idea that knowledge has agent-relative value. (On the distinction, see
Nagel 1986: 154.) More specifically, I shall work with the idea that knowledge as to
whether P has value for an individual concerned as to whether P. You have a concern
as to whether P when it becomes for you an open question whether P and you want
or need the matter to be settled. Such a concern might be enquiry-motivating; not
having a stance on the matter you are motivated to enquire into whether P out of
some practical or theoretical need or just out of curiosity. The question then is why,
with such a concern, one should want to know the truth rather than merely have a
true belief. One can, of course, also have a concern as to whether P when one already
believes that P or believes that not-P. For, even if one has a belief on the matter, the
truth of what one believes can become an issue to be addressed, especially if it comes
to seem that what one believes might be false. One might be prompted to wonder
because of a challenge or simply because, not having thought about the matter for a
while, one thinks afresh about it. Already having a belief on the matter does not make
otiose the consideration as to whether P. We can easily have occasion to evaluate
something we already believe. The question then is why it should matter whether we
have knowledge.

⁸ The translation is that of Alex Long in Sedley and Long (2011).

        ? 

Mere true beliefs, according to Plato, are susceptible to being undermined. They
can be undermined because our own reflections, observations, or social interactions
throw up challenges to them, or doubts concerning them. Drawing on Plato’s
example in Meno, suppose that, aiming to get to Larisa, I believe that I am on the
road to that town. I was told that it is the right road by someone who seemed to be a
reliable informant but someone else has persuasively cast doubt on that informant’s
reliability. I now have a concern as to whether I am on the right road. I do not know
what to think. I could persist along the road and hope to arrive at Larisa, or I could
make further enquiries, but, as things are, I do not know that this is the right road and
I now know that I do not know. If I am rational, I will not even believe that it is, since
my reason for believing is being challenged on what, for all I know, could be good
grounds. Contrast this with a situation in which I clearly remember taking this road
to Larisa last year. If in that situation the question were to arise as to why I take it to
be the right road I could readily answer that I went to Larisa by this very road last
year. The fact that I did so, which I have retained in memory, gives me a reason to
believe that it is the right road. By the same token it gives me reason to be
circumspect about any putative counter-evidence that might come to my attention.
Indeed, clearly remembering that I took this very road before, it should, and probably
would, take a lot to persuade me otherwise. Applied to the specific issue under
consideration about the agent-relative value of knowledge, Socrates’s suggestion
about the importance of tying true beliefs down seems to be heading in the right
direction.⁹ Even so, a caveat is in order.
We should beware of thinking that highlighting the importance of stability
commits us to explaining why knowledge matters solely in terms of its utility,
whether practical or theoretical.¹⁰ Stability of true belief is clearly useful: if true
beliefs are to be useful they need to be available to be exploited in case the need
arises; they cannot be exploited if they have slipped away. But this is only a part of the
story for, I claim, it is internal to a concern for the truth that it is not fully satisfied by
anything less than a stable grasp of it. Those wondering whether something is so
naturally want, and have good reason to want, more than a mere fleeting grasp of the
truth of the matter. Having reached the truth they want a secure grasp of it for at least
so long as their concern lasts. That is simply because if their true belief slips away
then their concern is no longer satisfied. Edward Craig suggests that stabilization of
true beliefs is not always important since ‘many beliefs are required for the guidance
of single, “one-off” actions under circumstances which will not recur, and once the
particular occasion is past there is no obvious value at all in their persistence’ (Craig
1990: 7). This is surely right, but given our focus on the agent-relative value of

⁹ For a similar, sympathetic reaction to Socrates’s response, see Williamson (2000: 78–9).
¹⁰ Duncan Pritchard in his contribution to Pritchard, Millar, and Haddock (2010: ch. 1, sec. 2) takes
Plato’s response to the problem of why knowledge is to be prized above mere true belief to be in terms of
the practical value of knowledge.
    

knowledge, the claim to defend should be only that stability of belief matters to one so
long as one is concerned as to the truth of the matter. The claim is not undermined by
the fact that many true beliefs will not need to be exploited after a certain point.
Those enquiring as to whether something is so have good reason to want to know
whether their enquiry has been successful. Enquiry of this sort is a task the aim of
which is to grasp the truth of some matter. To bring this task to successful comple-
tion enquirers must not only grasp the truth but also tell that they have done so. How
are they to do that if not by telling that they know the truth? And, of course, they
must know the truth in order to be able to tell that they know it. When things go well
they will be able to establish that they know it by reviewing how they came to know.
The reflective element here is crucial. It is not just knowing that they need in order to
complete enquiry but reflectively knowing—knowing while knowing that they know.
If the enquiry is into whether P, anything less than knowing that P or knowing that
not-P would not enable them to tell whether they have definitely grasped the truth. In
a whole range of matters in which knowledge is readily available this requirement is
just as readily met. Recall the case in which I look for my keys. I eventually look at the
telephone table and see, and thereby know, that they are there, but not only do
I recognize my keys as my keys, I recognize them as things I see to be my keys. So with
seeing that they are there comes knowing that I see, and in that way know, that they
are there.¹¹ The significance of reflective knowledge is that my enquiry does not
merely cease with the fixation of a belief as to the location of the keys. It ceases
because I know that the matter is settled—the truth securely grasped because known.
By contrast, with respect to a mere true belief, by definition I have no reason to think
it true and it comes with nothing else that makes sense of how the matter in hand can
be settled. The same applies to a mere justified true belief, if that is conceived in one
of the many ways that figure in mainstream epistemology. Even if I knew that I had,
in some such sense, a justified belief that something is so, this would not give me a
basis for taking it to be settled that it is so since, for all I know, I might be in a Gettier
situation—one in which I am relying upon a false assumption. If I am in such a
situation I might be able to tell that my belief is justified in the operative sense but
that would fall well short of settling that what I believe is true. (In Chapter 2, I argue
that justified belief in the sense in play in standard treatments of Gettier cases is not
the central notion of justified belief.)
Plausibly then, it is not just knowledge as to the location of my keys that I am after
but knowledge accompanied by knowledge that I know—reflective knowledge by
which I not only cease enquiry but know it to have been successfully completed.
Similar considerations apply when one already has a belief on some matter. On the
road to Larisa, if I know I am on the right road, and know how I know, then I shall
not be enquiring into whether it is the right road. If my belief had been merely true

¹¹ Epistemology needs an account of how this can be so. I present such an account in Chapter 5, sec. 5.
        ? 

I might have reached Larisa anyway, untroubled by doubts or challenges. All the
same, doubts and challenges could arise. Knowing that I know that I am on the right
road, in virtue of knowing how I know this, puts me in a good position to meet those
challenges and prevents my true belief from slipping away. From a practical point of
view this is a good position to be in, but it is more than that, because it is internal to a
concern as to the truth of a matter that one has reason to want both to know it and
know that one knows it.
It is not an implication of the view that only beliefs tied down in the manner
required for knowledge are stable.¹² Suppose it were possible to manipulate a person’s
psychology so that he or she comes to believe something and will continue to believe
it no matter what. Such a person would have a belief as stable as any that is tied to
knowledge; it would be tethered, and might be true, but the person would not thereby
have knowledge. The issue, though, is whether there is reason to want knowledge that
P if one is concerned as to whether it is true that P. True belief inculcated by non-
rational belief might be stable, but those concerned as to whether P have reason to
want to be in a position such that they can tell, that is know, that they have grasped
the truth of the matter. They can be in such a position if they do indeed know that P.
If they have a non-rational fixation, reflection upon it will not enable them to tell that
their concern as to the truth is satisfied.
Given what Plato had to say, it might be thought that the view I am defending
assumes that only belief tied down by evidence or other rational support can count as
knowledge, but this is not an implication of the view. Perceptual knowledge is a case
in point. You might know that the bird at which you are looking is a bullfinch simply
because you see it and recognize it to be a bullfinch from the way it looks, and not
because you have evidence or rational support in the form of reasons for the claim
that it is a bullfinch. (I defend such a view in Chapters 4 and 5.) It is not that you
would be haphazardly struck by the truth that it is a bullfinch in such a case. If you
know in this situation it is because you would have a general ability to recognize
bullfinches as bullfinches from the way they look and have exercised it on this
occasion. Reason would be at work in such a case, and, assuming you have the
conceptual resources that most of us have, you would have a reason to believe that it
is a bullfinch constituted by the truth that you see that it is. That truth could serve
both as a reason to think that there was a bullfinch in the garden and as an
explanation of how you knew it to be a bullfinch.
Finally, we should acknowledge that knowing something is not a guarantee that
the truth known will be securely grasped. Much that we come to know slips out of our
minds. This can easily happen when truths known are not routinely exploited. A lot
depends on how much the truths matter to us. None of this undermines the idea that

¹² This and the next two paragraphs are prompted by John Hyman’s criticisms of Plato’s view (2015:
199–200). It is instructive to consider whether they impact on my agent-relative conception of the value of
knowledge as to whether something is so.
    

when you are concerned about the truth of a matter you have reason to want to know
one way or the other.
I have used simple examples of perceptual knowledge, including that in which,
looking at my keys, I know where they are. It is not a routine task to extend the
general idea to other kinds of knowledge. This is something that merits close
consideration and I shall have more to say about it in the course of this book.
Among other things we need to make sense of the difference between, for instance,
telling that deer have passed from tracks on a path and coming to a true belief that
they have passed on the basis of the fact that they usually would have passed this spot
by this time of day. The trained gamekeeper can establish from the presence of tracks
that deer have passed and know that he has. He cannot do so on the basis of the deer’s
customary routine. Similarly, there is a difference between knowing that a family
member has arrived at her destination because she phones to tell you she has, and
coming to a true belief that she has arrived at her destination on the basis of the fact
that her means of travel were unlikely to have been delayed. The latter consideration
does not settle the matter; the phone call might do so. We should aim to make sense
of cases in which evidence is plausibly regarded as settling or establishing that
something is so. (See Chapter 2, sec. 5 and Chapter 8 for further discussion.)
It is not an implication of the view I have been defending that necessarily, if you
know then you know that you know. It is part of the view that knowledge, and indeed
reflective knowledge, is no mere aspiration. We often have knowledge and know that
we have it, which is why we can so often know that an enquiry has been successfully
completed and on that basis can truthfully vouch for the truth of the matter. For all
that, we might have less knowledge than some epistemologists think. Often the best
that we can hope to achieve are theories more or less well supported by evidence, but
not established as true. (See further Chapter 9, sec. 1.)
Problems arise in this area so long as we suppose that seeking the truth is something
distinct from seeking knowledge of the truth. It is undoubtedly the case that if
knowledge whether something is so matters, it is because the truth as to whether it is
so matters, but that is far short of establishing that knowledge matters because true
belief, merely as such, matters. It is not in dispute that if our search for the truth is
successful we shall have a true belief, but it does not follow that what we are after, or
should be after, if concerned as to the truth, is a true belief conceived merely as such.
True belief is a state that can obtain in virtue of satisfying minimal conditions for
believing a truth. It can be poorly grounded and unreasonable, fleeting and unstable.
Seeking true belief merely as such would be analogous to a climber’s wanting to hold
onto a rope merely as such, irrespective of whether she has a firm grip of it!
It might be thought that the view I have proposed about what is internal to enquiry
is open to question on the grounds that it generates a vicious regress.¹³ The view is

¹³ Jonathan Kvanvig (2011) raises this objection.

        ? 

that to be satisfied that one has completed the enquiry into whether P one needs to
know not just whether P but also whether one knows whether P. Suppose then that
I enquire into whether it is raining. I look from a window of my house and see, and in
that way know, that it is raining. To be satisfied that I have found out what I wanted
to find out, I have to know whether I found out that it is raining. I easily know this
because I know on looking, not just that it is raining, but also that I see, and in that
way know, that it is raining. But now it might seem that to be satisfied that I do
indeed know the latter—that I see, and in that way know, that it is raining—I need to
engage in further enquiry into whether I see, and in that way know, that it is raining,
and that this in turn necessitates yet further enquiry to determine whether that
enquiry has been successfully completed and so on, and so on. It seems that my
initial enquiry into whether it is raining could never be satisfied, because of the
endless need for further and further enquiries.
In the first place, it should be noted that the view I am defending does not entail
that all knowledge is acquired through enquiry, which is just as well since that view is
obviously false. Much of our knowledge is acquired irrespective of whether we have
undertaken any enquiry into the relevant matter. In the case in which I enquire into
whether it is raining, I establish that it is. In doing so I also come to know how I know
that it is raining, even though how I know this is not something I looked into and was
able to settle because I looked into it. I came to know on looking that I saw that, and
in that way knew that, it is raining, simply as an upshot of the enquiry into whether it
is raining.
A further problem might be thought to arise. The reason why enquiry into whether
P aims at reflective knowledge as to whether P is that reflective knowledge as to
whether P is necessary if one is to be satisfied that one has attained the truth as to
whether P. If having enquired into whether P, and acquired knowledge that P, a
question arose as to whether P and one did not know that one knew that P, for want
of knowledge as to how one knew it, one’s knowledge as to whether P is liable to be
unstable, and not apt to satisfy one’s concern for the truth. But now it might seem
that a similar problem arises for knowledge that is not the upshot of enquiry.
Returning to our example, it might seem that my belief that I see, and in that way
know, that it is raining, is liable to be unstable, unless I establish that I know that I see,
and in that way know, that it is raining. That requirement in turn then seems to
generate a need for further enquiry, and so we have a vicious regress, as before.
We need to recall what it is to have a concern as to the truth of a matter. A concern
as to whether P arises when it comes to be an open question for one whether P and
one wants or needs the matter to be settled. We imagined that on the road to Larisa
I become concerned as to whether I was on the right road. The concern makes sense
because it looked as if I might have been misinformed about the road. It becomes an
open question whether I am on the right road, and because I want to get to Larisa the
question is one that I need to address. In the case in which I find out that it is raining,
I might have been prompted to find out simply from curiosity. In any case, I have a
    

concern as to whether it is raining and I come to know, from what I see, that it is
raining. I also come to know on looking that I see, and in that way know, that it is
raining. The latter knowledge, though, is just as natural an upshot of my looking out
of the window as is the knowledge that it is raining. That is why, in such a
circumstance, in the absence of any reason to doubt that I see that it is raining, the
question whether I see that it is raining is not genuinely open. It cannot sensibly arise
for me, and my not having answered it is not liable to make unstable my assurance
that I see that it is raining.¹⁴ Whether one knows that one sees, and in that way
knows, that it is raining could come to be open to reasonable challenge and so
genuinely open. Such challenge, though, would be directed primarily at whether
one saw that it is raining. It would not presuppose that one saw that it was raining
and then raise a worry about whether, in addition, one knew that one saw that it was
raining.

4. Scepticism about conceptual enquiry

As I noted earlier, it was thought by many that a general account of knowledge would
take the form of a complete, reductive, conceptual analysis. Scepticism about the
possibility of such analysis, or about the likelihood of arriving at plausible analyses of
this sort for the topics of interest to philosophers, like knowledge, belief, justified
belief, intention, and so on, has fed into a more general scepticism about the value to
philosophy of reflection on concepts. The method I shall pursue in this book is
attentive to the character of the concept of knowledge, and in particular to its role in
our thinking, and our social interactions, but I do not envisage that enquiry into the
concept need or should aim at reductive analysis, complete or otherwise. It is of some
interest to see what such enquiry might be if it is not aimed at such analysis. I pick up
that theme in section 5 below.
The idea that philosophy should aim at conceptual analyses of the sort I have
described is notoriously subject to influential challenges. W. V. Quine in ‘Two
Dogmas of Empiricism’ (1953) famously attacked the view that there is a fundamen-
tal distinction between analytic or conceptual truths on the one hand, and synthetic
or factual truths on the other. Whether the attack succeeded in showing the view to
be false is debatable, but it undoubtedly contributed to a climate in which the very
idea of analytic or conceptual truths was held to be suspect, or at best uninteresting,
which in turn raised a question as to whether philosophy should be demarcated from
empirical enquiry on account of its preoccupation with such truths.

¹⁴ I do not imply that there is a necessary connection between knowing by seeing that it is raining and
knowing that one sees that it is raining; only that, given the conceptual abilities I actually have, the two go
together. The view that there is such a necessary connection is maintained in McDowell (2011). In
Chapter 5, sec. 5, I consider in more detail how we are able to acquire knowledge as to what one sees to
be so.
     

Scepticism about the scope of conceptual analysis was further fuelled by Hilary
Putnam’s work (notably his 1962/1975b and 1975a/1975b). Putnam argued plausibly
that we should not expect concepts of interest to scientists and philosophers to admit
of an analysis in line with the conceptual–analytical picture that I have been sketch-
ing. In a hugely influential treatment he argued that terms like ‘water’ and ‘gold’
stand for natural kinds and that it is the task of science to discover the essential
natures of those kinds. The essence of water, by Putnam’s account, is specified by its
microphysical structure, so water is essentially the substance H₂O. The philosophic-
ally significant conclusion was that we should not expect reflection on the concept of
water to reveal the nature of water—what makes water the substance that it is; only
empirical enquiry can do that. This seemed to open up the prospect that concepts of
interest to philosophers might pick out things—kinds, states, properties—the nature
of which would not be revealed by reflection on concepts. Hilary Kornblith (2002: 1)
could declare that ‘the subject-matter of epistemology is knowledge itself, not the
concept of knowledge’. It was never seriously in dispute that epistemology is con-
cerned with knowledge itself. The serious issue is whether philosophers interested in
realities should expect much help from reflection on concepts of those realities,
beyond some preliminary ground-clearing. There is a certain irony, then, in the
fact that Putnam’s own thinking accorded a central role to reflection on meanings
and can quite naturally be interpreted as according a corresponding role to reflection
on concepts.
Let it be granted that water is a natural kind, composed of H₂O molecules, and that
at least part of what makes it a natural kind is that samples of it are disposed to
behave in the same regular ways—ways that are explicable in terms of their sharing
the same essential nature (give or take impurities). Whether there is such a substance,
whether samples of it behave in the same ways, and whether their behaviour depends
on the molecules that compose it, are all matters for empirical enquiry. This is part
of Putnam’s picture, but the title of his most famous paper—‘The Meaning of
“Meaning” ’ (1975a/1975b)—and its content testify to the fact that he was interested
in the meanings of terms as we actually use them. A very natural way to understand
the thrust of his project is to take it to include issues about the concept of water,
conceived as a concept expressed by the term ‘water’ and thus grasped more or less
adequately by those who, more or less adequately, grasp the meaning of the term.
Because of assumptions he made about concepts, Putnam himself did not initially
think of his project in this way. He took grasp of a concept of an elm, for instance, to
amount to having a conception of what elms are (1975a/1975b: 226)—the particular
conception that sums up one’s own idea of what elms are and which might differ
from the conception that others have. While he acknowledged, indeed flagged up,
that the meaning of a term depends on social factors external to individuals, he
thought of concepts as being ‘in the head’. If, by contrast, we think of the concept of
water as something expressed by the term ‘water’ then we may view Putnam as
having shown us a way to think of concepts on which concepts are not identical with
    

conceptions in one’s head but are individuated by the meanings of the words used to
express them, and thus by the prevailing practice of using those words.¹⁵ With this
thinking in mind the core of Putnam’s reflections on water may be taken to
incorporate a theory of the concept that is expressed by the term ‘water’ as it is
and has been used in our linguistic communities. The theory includes the following
key claims:
(a) there is a substance S, which is a natural kind and satisfies our pre-scientific
conception of water as, among things, the stuff that fills rivers, lakes, and
oceans, that we drink to quench thirst, etc.;
(b) a substance is water if and only if it is the same substance as S;
(c) the term ‘water’, as we use it, and as it was used before the nature of water was
known as a result of the development of chemistry, expresses a concept that
correctly applies to S and only to S;
(d) it is possible that a substance should satisfy our shared pre-scientific concep-
tion of water and not be water.
This theory is open to challenge and no doubt, even at best, oversimplifies, but
(lo and behold) it illustrates a legitimate philosophical preoccupation with a
concept in which the aim is to become clearer about the metaphysical status of
the liquid to which the concept applies.¹⁶ It presupposes that there is an intimate
relationship—a certain harmony—between the metaphysical character of water
and the role of the concept in our thinking. We are to consider how the concept
would be applied or withdrawn in certain counterfactual circumstances, including
those on an imagined Twin Earth in which a substance that fits the pre-scientific
conception is found not to be composed of H₂O. The suggestion is that on Twin
Earth we would withhold or withdraw application of the concept to the substance
in question because it is not composed of H₂O, and thus not the same substance as
the substance in our familiar environment to which we apply the concept. More-
over, we are to think of this use of language as a manifestation of our grasp of the
meaning of the term ‘water’. We are to think of it as revelatory of the meaning of
the term ‘water’ and, we may now suppose, of the character of the concept of water.
We are to think of the concept as the concept of a natural kind and not that of
whatever fits the pre-scientific conception.
The theory is plausible only if, prior to the discovery that the water with which we
are familiar is composed of H₂O, the term ‘water’ was used in such a way that it is best
interpreted as correctly applying to, and only to, samples of a substance with a

¹⁵ Tyler Burge (1979) made a substantial contribution to this way of thinking. For some relevant
discussion, see Millar (2004: ch. 6).
¹⁶ Some critical discussion may be found in Pessin and Goldberg (1996). Putnam’s concession to Burge,
towards the end of his introduction to this volume of essays, lends some support to my reconstruction of
his thinking about natural kind terms.
     

uniform character that is explained by commonality of essential nature, rather than

functioning like the term ‘soil’ that correctly applies to stuffs with any of a wide range
of textures and compositions. Whether that condition is met is moot. My present
point is simply that it takes conceptual enquiry to settle the matter, hence the irony.
This conclusion is not at odds with the claim that there are limits to what reflection
on the concept of water can tell us about the liquid to which the concept applies. The
point is that whether the concept of water applies to a substance with a uniform
nature, give or take impurities, depends in part on the character of the concept and
that reflection on the metaphysical status of water and on the character of the concept
go hand in hand.
A further source of doubt about the value of conceptual enquiry is well represented
by Paul Churchland (1979; 1981). On his understanding of the notion of conceptual
truth, the sentences that supposedly express such truths are such that to deny them
would, as he puts it, be inconsistent with our understanding of the expressions that
make up those sentences. He assumed that the following would satisfy this condition
(1979: 47):

(1) Phlogiston is an elemental substance.

(2) Phlogiston forms compounds with other substances.
(3) Combustion and calxification both consist in the release of phlogiston from a
compound substance containing it.
(4) The release of phlogiston is induced by high temperatures.

He then asked what this has to do with truth, commenting as follows:

Let us put ourselves in the position of a proto-chemist who has just formulated his theory in
hopes of rendering intelligible the details of certain familiar physical transformations. Aside
from a few leading assumptions as to which substances are phlogiston compounds (woods,
metals), the set of assumptions (1)–(4) effectively exhausts our conception of phlogiston, our
understanding of the term ‘phlogiston’. These are sentences that introduce the term into
our general linguistic commerce, and it has no source of semantic identity, initially at least,
beyond these. In such a case the denial of one or more of these assumptions would indeed be
inconsistent with our understanding of the term ‘phlogiston’. And yet, though we ourselves
are the authors of these assumptions, we feel no temptation to claim that they are necessary
truths, that they are empirically irrefutable, that they are true solely in virtue of the meanings
of their terms. On the contrary, we are aware of the highly distinct possibility that one or
more, or even all of them, are false. And rightly so, for as we know, that is how in fact the
story ended. (Churchland 1979: 47)

We have here the idea that the meaning of a theoretical term, and correspondingly
the identity of a concept, can be fixed by a number of assumptions. The notion of a
theoretical term is liberal. A term is theoretical if its meaning is in part fixed by a
number of assumptions. With this conception in play it becomes conceivable that
falsehood can be incorporated within our concepts because the relevant meaning-fixing,
    

concept-individuating, assumptions can be false.¹⁷ Considerations along those lines

were the source of the idea that reflection on our concepts, at best, reveals folk theories—
clusters of platitudes that inform our employment of the concepts, but which might be
inadequate as theories of the realities that the platitudes are about. In Churchland’s own
work the outcome was eliminativism about our thinking about the mind—the view that
our ‘folk theory’ of mentality is ‘troubled or stagnant’ (1981: 75) because of its severely
limited explanatory and predictive success.¹⁸
The kind of thinking about concepts illustrated by Churchland has become
orthodoxy among strongly naturalistic philosophers. For instance, Jesse Prinz, writ-
ing on the emotions says:
Analyzing an emotion concept is assessing our beliefs about an emotion. . . . If I draw conclu-
sions about emotions when I engage in conceptual analysis, there is no guarantee that I will
arrive at a correct theory of what the emotions are. I may end up regurgitating some folk
psychological ‘platitude’ that turns out to be false. (2004: 29)

This is just the kind of attitude that informs Kornblith’s critique of the relevance of
conceptual analysis in epistemology. ‘Why’, he asks, ‘should our folk epistemological
notions be of any more interest to epistemologists than our folk chemical notions are
to chemists?’ (2002: 19). We should, though, be wary about the idea that thinking
about concepts of interest to philosophers is a matter of uncovering clusters of
assumptions that amount to a ‘folk theory’. Apart from anything else the very idea
that we have concepts individuated by clusters of assumptions is disputable. An
obvious worry is that it too finely individuates concepts, since on this understanding
changes in the individuating theory would, implausibly, amount to changes in
concept, that is, a shift from one concept to another. Arguably, it also misconceives
the relation between concepts and conceptions. Dictionary definitions of terms often
describe that to which the term correctly applies. The Shorter Oxford English
Dictionary, for instance, has an entry for ‘water’ that reads as follows:
The liquid (in its pure form transparent, colourless, tasteless, and odourless) which forms the
main constituent of seas, lakes, rivers, and rain, and is put to many domestic and industrial
uses: it is a compound of hydrogen and oxygen (formula H₂O), and was formerly regarded as
one of the four elements.

The description here includes clearly contingent features of water, including that it is
put to many domestic and industrial uses. It articulates a conception of water, but
there is nothing about its role that licenses identifying the concept of water with the
conception articulated. Rather, the description fixes the meaning of ‘water’ simply by

¹⁷ For another discussion sceptical of conceptual analysis, and the metaphysical interest of ‘folk’
psychology, see Stich (1983: ch. 2). For a contrasting view, see Frank Jackson (1998).
¹⁸ Finn Spicer (2008) makes a case for thinking that our ordinary conception of knowledge is incon-
sistent. I am not persuaded.
 ’    

telling us what the term applies to.¹⁹ Entries for ‘water’ that differ somewhat from this
entry might also serve to convey the meaning of the term by telling us, albeit
somewhat differently, what it correctly applies to. To count as grasping the meaning
of the term one needs to have some conception of what it applies to, but it is not
necessary that everyone with such a grasp works with exactly the same conception,
and shifts in conception over time need not amount to changes in concept. Even if
some concepts are individuated by clusters of assumptions, it would take argument to
show that this is true of any particular concept, and in any case there is no reason to
assume that conceptual enquiry in general aims at uncovering concept-individuating
assumptions.
In the next section I argue that, notwithstanding the history that I have all too
briefly sketched, there is a significant role for conceptual enquiry in philosophy in
general and epistemology in particular.

5. Philosophy’s interest in concepts

It is striking that influential figures of twentieth-century philosophy, who sought to
illuminate philosophical topics through reflection on meanings and concepts,
showed little if any interest either in giving complete reductive analyses of concepts
in the traditional manner or in explicating meanings or concepts in terms of clusters
of assumptions.
Gilbert Ryle is standardly regarded as a logical or analytical or philosophical
behaviourist—someone who thought that our statements about the mental could
be given a complete reductive analysis in terms of dispositions to behaviour.²⁰ Yet,
Ryle showed no interest in such analyses but focused on identifying and diagnosing
metaphysical muddles through reflection on how we think about aspects of our
mentality, when not driven by philosophical considerations. He thought that atten-
tion to our ordinary (non-philosophical) thinking about the mind would reveal that
the concept of the mind that informs that ordinary thinking is not that of an
Aristotelian substance, about which it would make sense to ask whether it is or is
not identical with anything bodily. Rather, we are to think of the mind as in large
measure a distinctive set of capacities, dispositions, and tendencies. In thinking about
belief, for instance, we are to imagine someone who, skating on a frozen pond,
believes that the ice is dangerously thin at a certain dark patch (Ryle 1949/1963: ch. V,
sec. 3, 129–30). Assuming that she has a suitable background of beliefs and other
attitudes, including a desire not to fall through, and a concern that others should not
do so, she will be disposed to react accordingly by, among other things, seeking to
avoid the thin patch, dwelling in imagination on the danger, warning others should

¹⁹ Compare Wiggins (2007), especially sec. 4.

²⁰ See, for example, Armstrong (1968: 54); Braddon-Mitchell and Jackson (1996: 29, 40). For a more
nuanced, and more accurate, discussion of Ryle’s ‘behaviourism’, see Byrne (1994).
    

the need arise, and so forth. Ryle’s aim, I take it, was to illustrate the general point
that it belongs to a belief about one’s surroundings that it has the potential to impact
on how one behaves in relation to those surroundings and on further thoughts and
concerns that one might have that relate to those surroundings. This is now widely
accepted.
Thinking along those lines is an exercise of the imagination aimed at becoming
clearer about what it would be to have the belief in question. It would badly
misrepresent the procedure to describe it as concerned solely with the concept of
belief, as opposed what belief is. A presumption of the enterprise is that when we
think about beliefs we are not thinking of fictions. While the mere fact that we have a
concept of belief does not guarantee that it correctly applies to anything, we have no
reason to suppose that it does not have correct application. On the contrary, we
know a great deal about what we, and others, believe. We have a sense of when we
have or lack an understanding of what people say they think, of how to seek
clarification of what people think, and so forth. There is, then, a feasible project of
considering what beliefs must be given, not just what we say or imply about beliefs,
but what we already know about beliefs and how we deal with them. The task is
interpretative because, though we already understand ascriptions of belief at some
level, we seek a better understanding of what it is to believe something—one that will
enable us to survey how believing something links up with, for instance, other
beliefs, intentional states of other sorts, and actions. Donald Davidson’s ideas are
pertinent at this point.
Davidson’s thinking about understanding others is governed by a conception of
the individuation of intentional states that is well expressed in the following
passage:
Individual beliefs, intentions, doubts and desires owe their identities in part to their position in
a large network of further attitudes: the character of a given belief depends on endless other
beliefs; beliefs have the role they do because of their relations to desires and intentions and
perceptions. These relations among the attitudes are essentially logical: the content of an
attitude cannot be divorced from what it entails and what is entailed by it. This places a
normative constraint on the correct attribution of attitudes: since an attitude is in part
identified by its logical relations, the pattern of attitudes in an individual must exhibit a
large degree of coherence. This does not, of course, mean that people may not be irrational.
But the possibility of irrationality depends on a background of rationality; to imagine a totally
irrational animal is to imagine an animal without thoughts.
(Davidson 1994: 232. Similar remarks occur in Davidson 1975/1984: 159; 1982/2001: 99.)²¹

The point here is that a charitable construal of how people are thinking or acting or
feeling—one in the light of which it is intelligible that they should think, act, or feel as
they do—is no mere imposition, since it is required by the very nature of intentional,

²¹ I discuss this passage and others from Davidson in ch. 1 of Millar (2004).
 ’    

concept-implicating, states. Moreover, the role of perception in acquiring beliefs and

knowledge brings out that there is no divorcing what we make of the lives of others
from how they relate to their surroundings. What obtains or occurs in the natural
and social environment we share with others contributes to our reading of what they
think about that environment and how they react to it. (Compare Chapter 5, sec. 7 on
Quine and McDowell.) We understand why a woman rushes towards a child because
we see that the child is about to walk onto a busy road and realize that she sees this
and wants to make the child safe. We rely on our own knowledge of what is going on,
including our knowledge of what the woman can see, and in that way know, to be
happening.
Philosophical enquiry into belief is served by consideration of what we make of
ascriptions, and avowals, and evaluations of belief in the light of what we already
know about belief or, more precisely, about the range of states that beliefs can be—
convictions, opinions, expectations, degrees of confidence, and so forth. If the aim
were simply to devise a conception that will fit with our ordinary judgements (so-
called intuitions) about belief we could reasonably wonder why a conception that
meets this condition should be expected to be true. (This theme is taken up again in
Chapter 9, sec. 1.) The conception sought is rather one that will fit with what we
know. What we know encompasses not just facts about what people actually believe
and how they think but also considerations pertaining to what is and is not intelli-
gible. We might well think about imaginary cases, but the cases we select must reflect
what we know about the circumstances in which, and the grounds on which, people
acquire and give up beliefs and exploit those beliefs in further thought or action. The
work goes into considering what believing must amount to, given those known
truths. To engage in such work we need appropriate conceptual competence, but
the aim is to go beyond the understanding that such competence gives us with a view
to reaching a deeper understanding of belief. In all of this we are fallible. In advance
of enquiry into some area we should not assume that our ordinary thinking in that
area is perfectly in order, but nor is there any need to presume that it is likely to be
seriously at fault unless good reason emerges for supposing that it is. (I take up the
idea that philosophy must be guided by what we know in further reflections on
method in Chapter 9.)
In the light of these considerations it does not look plausible that philosophical
enquiry into the nature of belief is a matter of reflectively exploiting one’s grasp of the
concept of belief in a purely a priori manner. This really should be no surprise. The
kind of exercise of the imagination that I have been describing needs materials to
work with. These can only come through experience of acquiring beliefs, exploiting
beliefs, giving up beliefs, making sense of what people do in the light of their beliefs,
and so forth. Drawing on this experience, the aim, to repeat, is to answer the
question, ‘What is it to believe something?’ by devising a conception of belief that
makes sense of what we already know about belief, and enable us to articulate a
deeper understanding of what belief is.
    

I have been using belief merely as an illustrative example. The general methodo-
logical approach applies as much to knowledge as it does to belief, and to many other
states of interest in philosophy. A key assumption is that reflection on concepts is
inextricable from engagement with the realities that those concepts enable us to think
about. In all of this I have kept in mind how one might expand on these remarks from
J. L. Austin:
When we examine what we should say when, what words we should use in what situations, we
are looking again not merely at words (or ‘meanings’, whatever they may be) but also at the
realities we use the words to talk about: we are using a sharpened awareness of words to
sharpen our perception of, though not as the final arbiter of, the phenomena.
(1956–7/1979: 182)

Austin would have thought it bizarre to suppose that enquiring into the concept of
knowledge should be contrasted with enquiring into what knowledge is. Interest-
ingly, there is nothing in the quoted passage that is inconsistent with the idea that
looking into concepts should form part of an enquiry that takes us beyond the
understanding we have simply in virtue of grasping those concepts.
The considerations I have been setting out are especially pertinent to epistemology,
for this is an area where reflections on what knowledge is, what justified belief is,
what constitutes evidence, what reasons can be, and so forth, often seem to occupy a
sphere that is quite detached from ordinary ways of thinking. To those who think you
can do metaphysics or cognitive psychology with scant regard for ordinary ways of
thinking this will seem to be no problem at all. I am suggesting that philosophical
enquiry has an interpretative dimension directed at understanding what knowledge
is in the light of what we know and say about knowledge and about our ways of
thinking and talking about knowledge. To proceed in this way is not to suppose that
epistemology can never be revisionary. Though knowledge is not an elusive com-
modity its scope, I suspect, is narrower than it is widely thought to be, because what
we can establish or settle is narrower than what many take the scope of knowledge to
be, but we should not lose sight of the fact that we know quite a lot and that we
already know something about knowledge itself, just as we do about belief, intention,
and other matters of philosophical interest. Much of what we know about knowledge
is bound up with what we know about ways to tell whether this or that is so. I shall
have this before my mind at every stage in this book.
 2
Justified Belief, Reasons,
and Evidence

1. Justified belief
A major strand of contemporary epistemology assumes that justified belief must be
explicated independently of the concept of knowledge. In this chapter, I outline a way
to think of justified belief on which knowledge is explanatorily prior to justified belief.
In Chapter 1, I suggested that in thinking philosophically we should have due
regard to the ways in which relevant concepts figure in ordinary (non-philosophical)
thinking. This goes for enquiry into justified belief as much as it does for enquiry into
knowledge. Stewart Cohen has expressed some scepticism about whether it makes
sense to proceed in this way. This method, he thinks, is problematic because while
‘the concept of knowledge plays a central role in the way we view ourselves and our
relation to the world’ (1995: 114) the same is not true of the concept of justified belief.
He writes:
. . . ‘justified’ is a term that we do not ordinarily apply to beliefs. So there is no such thing as
what we mean when we ordinarily say of people that they have justified beliefs. There is no
ordinarily used expression ‘justified belief ’. The description of beliefs as justified is something
that occurs primarily in conversations among working epistemologists. In ordinary parlance,
we describe beliefs (or people in virtue of believing) with terms like ‘reasonable’, ‘rational’, or
‘reliable’. We ask ‘Why does Jones believe that . . . ?’ We inquire about what evidence or reasons
a person has. We wonder how likely it is that what a person believes (or says) is true. But it is
certainly rare, outside of philosophical contexts, for a speaker to use ‘justified’ in connection
with a belief (or a person in virtue of believing). (Cohen 1995: 116)

The situation is undoubtedly complicated by the range of very different conceptions

of justified belief in contemporary epistemological debate. As I noted at the begin-
ning of Chapter 1, justified belief is conceived in terms of reasons, or evidence, or in
terms of some broader notion of right or entitled or rational believing, and even in
terms of being formed by a mode of belief-formation that would reliably yield true
beliefs. Some of these are undoubtedly specialized notions, introduced for theoretical
purposes. Nonetheless, whatever may be said of the term ‘justified belief ’, there is in
our ordinary (non-philosophical) thinking a concept that the term may reasonably
be taken to express. It is that of believing something for an adequate reason.

Knowing By Perceiving. Alan Millar, Oxford University Press (2019). © Alan Millar.
DOI: 10.1093/oso/9780198755692.003.0002
  , ,  

Questions like ‘Do you have reason to think that true?’, requests like, ‘Give me a
reason to think that true?’, and comments along the lines of, ‘That is no reason to
suppose that . . . ’ are widely understood. They concern normative reasons for belief,
that is, reasons to think something true. A reason to think it true that P favours or
counts towards its being true that P. In ordinary parlance we speak of there being
some reason to think that something is true, where that means little more than that
there is some ground for suspecting it to be true. The notion of an adequate reason to
believe something is more restrictive. An adequate reason to believe that P is a
consideration that entails, or makes it highly likely, that P. To be justified in believing
something, we may say, amounts to having an adequate reason to believe it and
believing it for that reason. This gives us the reasons-conception of justified belief. I am
far from confident that the term ‘justified’ is not actually used in keeping with this
conception. It does not strike me as odd that one should be asked to justify a claim
that one has made, where that is understood to be a request to back up the claim, and
so to give reasons for thinking it true. To comply would be to represent the claim, and
thus any belief it expresses, to be justified. That said, nothing for which I shall be
arguing turns on whether the term ‘justified’ is routinely used in this way. What
matters is that the concept does figure in our thinking. There is a relatively familiar
notion of believing something for an adequate reason.
Normative reasons to believe something are constituted by truths or facts.¹ In
presenting our reasons for believing something we may, of course, present something
false, but the aim is clearly to present truths. If I seem to see a friend of mine walking
along the street, I might think that this friend is back in town, but if the person I see is
not my friend but someone else, then I lack the reason I thought I had for believing
that he is back in town. Even in that circumstance there is a sense in which I believed
for a reason, but the reason in this case is my reason for believing that my friend is
back in town—a motivating or explanatory reason.² Note that a belief can be
overdetermined by motivating reasons. It can be that one believes that P for the
reason that, as one supposes, Q, and also that one believes that P for the distinct
reason that, as one supposes, R.

¹ This claim is regarded as controversial in the literature on epistemology. I take it to be entirely in

keeping with common sense. There is no good reason to be revisionary about this. So long as we attend to
the distinction, introduced in this paragraph, between normative reasons and motivating reasons, we can
make sense of the evident fact that people can believe for reasons that are constituted by propositions that
are false. The claim that reasons for action are constituted by truths or facts is widely accepted in the
literature on practical reason and on normativity more generally. See, for instance, Skorupski (2010), Parfit
(2011), and Scanlon (2014). (Note that for my purposes truths may be equated with facts.)
² The distinction between normative and motivating reasons is familiar in the literature on reasons for
action, but less so in the literature on epistemology. I use it in connection with both believing and acting in
Understanding People (Millar 2004). In the sense intended a reason for which one believes something, thus
a motivating reason, need not be a reason on account of which one comes to believe it. It can be a reason
that sustains a belief as opposed to bringing it about. This turns out to be important in connection with
perceptual justification. (See Chapter 5.) I shall persist in using ‘motivating’ in preference to ‘explanatory’
since the latter might be thought to apply to any reason why something obtains or happens.
   

Let us look more closely at what it is to believe something for a reason.³ When one
believes something for a reason, the reason might or might not be a reason that
favours believing that thing, thus a normative reason so to believe. In either case the
reason must be constituted by considerations in the light of which it is intelligible—
explicable—that one should have believed as one does. Beliefs (believings) can be
explained and, in some cases, shown to be justified in the light of the considerations
that constitute the subject’s reason for holding them. Explanations in terms of the
subject’s reasons—rationalizing explanations—advert to how things seemed from the
subject’s point of view. The extent to which a belief or action is intelligible depends
on how intelligible it is that the subject should have treated the relevant reasons as
normative reasons for that belief or action. The attempt to explain is therefore bound
up with evaluation of the subject’s reasons.
It is readily intelligible that someone should believe something if his or her reason
for believing this thing is clearly an adequate reason so to believe. Indeed, it is a
condition of a reason’s being an adequate reason for believing something that
believing it for that reason, and so believing it in the light of the considerations
that constitute it, should adequately explain one’s believing it. You might believe it
has rained recently for the reason that the ground and other surfaces are wet (in the
way they would be after rain). In this case the fact that you believe it has rained
recently for that reason suffices to explain your forming this belief and the explan-
ation works from both first-person and third-person standpoints. If I believe that it
has rained recently for the reason that the ground and other surfaces are wet then
I can make it intelligible to myself that I believe that it has rained recently in view of
the fact that the belief is one for which I have this adequate reason.
We can, of course, believe things for bad reasons. To explore this it helps to work
with the notion of an adequate reason-giving relation. I shall say that considerations,
Q, R, . . . , stand in an adequate reason-giving relation to the proposition that P
provided that, were they true, they would constitute an adequate reason to believe
that P. There will be occasions when a subject believes something for a reason, and
the considerations constituting the reason stand in a reason-giving relation to what
the subject believes, but some element in those considerations that is essential to their
being reason-giving is false. That gives us one type of bad reason. There is another
type since it is plausible that one can believe for a reason that is bad because what
constitutes it does not stand in an adequate reason-giving relation to the thing
believed. Suppose that I took a dim view of some colleague, believing that he shirks
his responsibilities. My reason for believing this was that, as I supposed, he refuses to
take on certain jobs. The truth of the matter is that he did refuse to take on those jobs
when asked to, not because he generally shirks responsibilities but because, unknown
to me, he already had more than enough on his plate. My reaction was due to

³ Here I return to a central theme of Millar (2004: ch. 7).

  , ,  

prejudice on my part against this colleague—I too readily took his refusal to be
indicative of a general attitude towards taking on responsibilities. In such a case, it is
still true that I believed what I did about this colleague for a reason, even though what
constituted my reason did not stand in an adequate reason-giving relation to what
I believed. Too easily might it have been the case that even though he refused to take on
the jobs he is not one who shirks responsibilities. So, the fact that I believed as I did for
this reason does not suffice to explain why I believed as I did. Still, there is a sense in
which the basis of my belief rationalizes the belief, in that to understand why it was
formed you have to take into account a consideration in the light of which it was
formed—a consideration that I treated as an adequate (thus normative) reason to
believe as I did. It is not that the belief was formed solely because of factors that have
nothing to do with the operation of the intellect. It is just that my prejudice distorted
the operation in this case. The rationalization is, as we might say, messy because the
basis of my belief did not stand in an adequate reason-giving relation to what I believed.
Such cases pose a problem for a tempting account of believing for a reason,
according to which believing something for a reason is always a matter of believing
it for what would be an adequate reason to believe it if the considerations constituting
it were true. Such an account is falsified by cases of messy rationalization. Yet, there
are constraints on believing for a reason; not just anything could in principle be one’s
reason for believing something, no matter how irrelevant to the justification for
believing that thing. There is a linking principle to the effect that when one believes
that P for the reason that, as one supposes, Q, R, . . . , one treats the considerations
that Q, R, . . . as giving one an adequate reason to believe that P.⁴ This need not
involve bringing to mind that the considerations give one a reason to believe that P.
For subjects who have appropriate reflective capacities it is a matter of, for instance,
being prepared to explain why one thinks that P in terms of the reason-constituting
considerations, and being prepared to adduce those considerations by way of justi-
fying believing that P. Without the linking principle there would be no accounting for
what may count as a motivating reason.
When one regards something as an adequate reason to believe something else one
exploits one’s understanding of the significance of what one takes to be the truth or
truths constituting the reason. In messy cases the understanding one exploits is a
misunderstanding. Nonetheless, in such cases what one treats as an adequate reason
to believe must comprise considerations in the light of which it is intelligible that one
should have formed or maintained the belief on their basis, and so intelligible that

⁴ Stephen Darwall makes a similar point in connection with motivating reasons for action as follows:
Something may be somebody’s reason for having acted without having been a reason for
him so to have acted . . . , but it must nonetheless be a consideration that he regarded (or
perhaps would regard under certain conditions) as a reason for him so to act. What
characterizes explanation of action in terms of the agent’s reasons is that it explains it as
an expression of the agent’s own conception of what reasons there were for him to act.
(Darwall 1983: 32)
   

one should have treated them as giving one an adequate reason so to believe. If the
case is one in which the considerations constitute an adequate reason to believe then
it will be readily intelligible that the subject should have formed or maintained the
belief in their light. But people are fallible; they sometimes reason poorly and can be
subject to various kinds of motivational bias. So sometimes they believe things for
bad reasons. Still, if it can so much as seem to a subject that certain considerations
constitute a reason to believe something then these, along with any other relevant
factors, should enable us to understand how from the subject’s perspective they could
have seemed to provide an adequate reason to believe the thing in question.
My believing that my colleague shirks responsibilities on the grounds that he has
refused to take on certain jobs is intelligible because, given my standpoint, it is
intelligible that I should have taken his refusal to supply an adequate reason so to
believe. After all, the fact that he refused to take on the responsibilities is relevant to
the claim that he shirks responsibilities in that such a fact could be part of evidence
supporting such a claim. That I treated it as an adequate reason to believe that he
generally shirks responsibilities is explained by the prejudice that took up the slack
between what my reason shows and what I believed. While subject to the prejudice
my own capacity to evaluate my belief was impaired if only to the extent of reducing
my willingness to assess whether I had good reason to believe as I did. If I had
reflected, it ought to have been obvious that there could be an explanation for my
colleague not taking on the job that does not involve the assumption that he generally
shirks responsibilities. In this case we still have a rationalization, albeit a messy
rationalization, of my belief; given my prejudice, it is intelligible that I should have
taken the consideration constituting my reason to be an adequate reason to believe as
I did. In acknowledging that I would not have held the belief but for the prejudice,
I am committed to supposing that any reason I had for holding it was inadequate.
Self-ascription of a belief is inextricable from endorsement of its content.⁵ It is
contrary to reason to acknowledge that you have no adequate reason to endorse a
belief ’s content while retaining the belief. (There is no tension here with the fact that
we can submit our beliefs to scrutiny. When we seriously evaluate a belief we hold it
in abeyance. It becomes an open question whether we should continue to hold it, so
that, pending the outcome of the evaluation, we do not at that point endorse its
content.)
The topic of believing for bad reasons serves to highlight that when we believe
something for a reason there is a sense in which we believe it in the light of that
reason. Rightly or wrongly we treat what constitutes the reason as providing us with
an adequate reason to believe the thing in question. This has bearing on the
discussion of evidence in section 3 below.

⁵ I explore related matters in Millar (2004: chs. 4 and 5.7). See also Chapter 7, sec. 4 of this volume.
  , ,  

That normative reasons are constituted by truths or facts is reflected in the factivity
of the notion of believing for the reason that. . . . The claim that X believes that P for
the reason that Q entails both that the reason in question is a normative reason to
believe that P and that it is true that Q. If I accept that X believes that P for the reason
that Q, I incur a commitment to accepting that Q. We may, of course, wish to
attribute a motivating reason without incurring such a commitment. We may do so
by saying that X believed that P because he believed that Q, but some care is needed
here. It should be not inferred that motivating reasons are constituted by beliefs in
the sense of believings. They are constituted by propositions that the subject, rightly
or wrongly, takes to be true and treats as providing an adequate reason to believe that
P. We may also avoid the commitment by saying that X believed that P because, as he
supposed, Q, a manner of speaking that I have adopted in this chapter.⁶
Jennifer Hornsby (2008) has an objection to invoking a distinction between
motivating reasons and normative reasons. Adapting one of her examples on reasons
for action, suppose that Ann informs me that a named café where we are to meet is at
a certain spot at the west end of George Street. We are to meet in the café that Ann
named, but it is not in George Street but elsewhere. It seems that I now have a reason
to head for the west end of George Street—a reason constituted by the, as it happens,
false consideration that the café is there. It also seems that I have a reason to believe
that we are to meet near to the west end of Princes Street, since the relevant location
in George Street is one that I know to be close to the west end of Princes Street. This
reason is in part constituted by the false proposition that the café is at the west end of
George Street.
Hornsby stresses that, as she understands it, having a reason to Φ is not the same
as there being a reason to Φ. From the perspective that I have been sketching, a
reason for me to Φ is a normative reason, and normative reasons are constituted by
truths, but even if I lacked such a reason to go the west end of George Street in the
situation envisaged, there is still a sense in which, in keeping with the spirit of
Hornsby’s treatment of the example, I might be said to have a reason to do this in
that I have a reason of the sort that I am calling a motivating reason. Hornsby,
though, thinks that it is misleading to classify the reason as motivating by way of
distinguishing it from a normative reason. The rationale she gives for this is that
‘normative questions, about the justification of action, don’t lapse when one adverts
to what the agent believes in saying what reason they had’ (Hornsby 2008: 249–50).
In relation to the example, and extending the idea to reasons one has for a belief,
Hornsby’s point is that there is a normative dimension to my reason for heading to

⁶ Jonathan Dancy (2008: 270ff. and 2000: 132) thinks that one may do something for the reason that P
even when not-P. Dancy (e.g., at 2008: 270) rightly resists the suggestion that when it is false that P, and one
Φs in the light of the consideration that P, one’s reason is that one believes that P, but this does not support
the claim that one may Φ for the reason that P when not-P. For critical discussion of Dancy in relation to
the present issue, see McDowell (2013).
    -  

the west end of George Street, and to my reason for thinking that we are to meet near
the west end of Princes Street, and this normative dimension is shared by reasons
that are wholly constituted by truths, but not shared by, for instance, the reason why
a bridge collapsed. As Hornsby views the matter, invoking the notion of a motivating
reason, and placing such reasons in a different category from that of normative
reasons, threatens to assimilate motivating reasons to the kind of reason-why that
would figure in an explanation of why the bridge collapsed. But there is no real threat
here. We should resist the assimilation, but doing so is compatible with acknowledg-
ing that motivating reasons and normative reasons are in different categories,
notwithstanding that what constitutes a motivating reason can constitute a norma-
tive reason, and vice versa. My reason for heading to the west end of George Street,
and for believing that we are to meet near to the west end of Princes Street, are
motivating reasons, but not normative reasons. There is a normative dimension to
motivating reasons, but this, I suggest, is brought out by the connection already noted
between what it is to be a motivating reason and what it is to be a normative reason.
A motivating reason for one’s Φing/believing that P is constituted by a consideration
that one treats as providing a normative reason to Φ/believe that P and that,
moreover, does provide something that could intelligibly seem to the subject to
constitute such a reason. That connection is not such as to make it true that what
constitutes a motivating reason always constitutes a normative reason, but nor does it
assimilate motivating reasons to the kind of reason that figures in the explanation of
the bridge’s collapse.

2. ‘Justified’ belief in Gettier cases and well-founded

belief
In contemporary epistemology, justified belief has been of interest largely because it
has been widely thought that knowledge should be analyzed in terms of true justified
belief, even if other conditions need to be added to yield a complete analysis. The
quest for an analysis that would be reductive dictated that justified belief should itself
be understood independently of knowledge. We are supposed to pin down what
justified belief is in terms that do not bring into play the concept of knowledge, so
that even if we had no idea what knowledge is we could still understand what justified
belief is. Such a stance might seem attractive if one thought that we can understand
what is really central to enquiry and rationality and, more generally, to our intellec-
tual lives, without invoking the concept of knowledge. The stance taken here is very
far removed from this. I claim that justified belief, in the central sense, is well-
founded belief and that well-founded belief is founded on something one knows.
Knowledge, on this view, is prior in the order of understanding to justification, and
justification is prior in the order of understanding to another standing, which I shall
call reasonableness.
  , ,  

An important part of the story relates to the discussion of Gettier cases. Many of
these cases are of assumption-based beliefs. These are beliefs we form or retain on the
basis of assumptions that we accept and treat as a reason to believe as we do. We
might be prompted to believe by coming to accept the assumptions, or we might have
accepted the assumptions for some time and be led to believe on bringing the
assumptions to mind. A standing belief, however formed, can also be based on
certain assumptions when the belief is sustained by those assumptions.⁷
Timothy Williamson (2007: 183) has this Gettier case, which is of a classic type:
A clever bookseller fakes evidence which appears to show conclusively that a particular book
once belonged to Virginia Woolf; convinced, Orlando pays a considerable sum for the book.
He has a justified false belief that this book of his once belonged to Virginia Woolf. On that
basis alone, he forms the existential belief that he owns a book which once belonged to Virginia
Woolf. The latter belief is in fact true, because another of his books in fact once belonged to her,
although he does not associate that one with her in any way.

Cases like this are directed at analyses of knowledge on which it is assumed that there
is some way of being justified in believing that P such that one knows that P if and
only if it is true that P, and one is in that way justified in believing that P. In the given
case Orlando certainly relies on a false assumption and certainly does not know that
he owns a book that once belonged to Virginia Woolf. The standard reaction has it
that this belief is justified and therefore that having a justified true belief is not
sufficient for knowledge.⁸ If cases are to be effective in supporting such a view they
must be cases in which the subject has a true belief that is justified in the right way,
that is, in a way that is adequate to satisfying the supposed justification requirement
for knowledge. This raises problems for standard treatments.
First, those treatments assume that a belief can be justified even if based on
assumptions some of which are false. A belief so based can hardly be well-founded
in the sense of having a sound basis, since a sound basis is provided only by adequate
(normative) reasons and such reasons are constituted only by truths.⁹ Of course, we
need to take account of the difference between motivating and normative reasons.
One’s reason for believing something might not be constituted by a truth. This is so
in Orlando’s case, yet though his belief is not well-founded, and in that sense justified,
it could be that he had more than an excuse for believing that the book he bought
once belonged to Virginia Woolf. One’s belief that P can be reasonable in cases in
which one is misled and one’s being misled is not because of lack of due cognitive

⁷ See Armstrong (1973: 79–85) for discussion of sustaining causes of belief.

⁸ Williamson is not concerned to use the example for the usual purpose. He aims to expose lacunae in
standard treatments of such cases.
⁹ This strikes me as an entirely natural way to think of well-founded belief. When people stoutly claim
that a certain view is without foundation, they are denying either that the view is supported by truths or
facts, or that truths or facts adduced in support of the view provide an adequate reason to accept it. For a
contrasting use of the term ‘well-founded’, see Conee and Feldman (2004: 38–9, 93ff.).
Another random document with
no related content on Scribd:
 The Project Gutenberg eBook of
Kulturgeschichte der Deutschen im Mittelalter
This ebook is for the use of anyone anywhere in the United States
and most other parts of the world at no cost and with almost no
restrictions whatsoever. You may copy it, give it away or re-use it
under the terms of the Project Gutenberg License included with this
ebook or online at www.gutenberg.org. If you are not located in the
United States, you will have to check the laws of the country where
you are located before using this eBook.

Title: Kulturgeschichte der Deutschen im Mittelalter

Author: Georg Steinhausen

Release date: November 18, 2023 [eBook #72163]

Language: German

Original publication: Leipzig: Verlag von Quelle & Meyer, 1916

Credits: The Online Distributed Proofreading Team at

https://www.pgdp.net

*** START OF THE PROJECT GUTENBERG EBOOK

KULTURGESCHICHTE DER DEUTSCHEN IM MITTELALTER ***
 Anmerkungen zur Transkription
Das Original ist in Fraktur gesetzt. Im Original gesperrter Text ist so
ausgezeichnet. Im Original in Antiqua gesetzter Text ist so markiert.
Weitere Anmerkungen zur Transkription befinden sich am Ende des Buches.
 Wissenschaft und Bildung
Einzeldarstellungen aus allen Gebieten des Wissens
Im Umfange von 150–180 Seiten
Geh. 1 M. · In Leinenband 1.25 M.
Die Sammlung bringt aus der Feder unserer berufensten Gelehrten
in anregender Darstellung und systematischer Vollständigkeit die
Ergebnisse wissenschaftlicher Forschung aus allen
Wissensgebieten. :: :: :: ::
Sie will den Leser schnell und mühelos, ohne Fachkenntnisse
vorauszusetzen, in das Verständnis aktueller wissenschaftlicher
Fragen einführen, ihn in ständiger Fühlung mit den Fortschritten der
Wissenschaft halten und ihm so ermöglichen, seinen Bildungskreis
zu erweitern, vorhandene Kenntnisse zu vertiefen, sowie neue
Anregungen für die berufliche Tätigkeit zu gewinnen. Die Sammlung
»Wissenschaft und Bildung« will nicht nur dem Laien eine
belehrende und unterhaltende Lektüre, dem Fachmann eine
bequeme Zusammenfassung, sondern auch dem Gelehrten ein
geeignetes Orientierungsmittel sein, der gern zu einer
gemeinverständlichen Darstellung greift, um sich in Kürze über ein
seiner Forschung ferner liegendes Gebiet zu unterrichten. Der
weitere Ausbau der Sammlung wird planmäßig durchgeführt.
Abbildungen werden den in sich abgeschlossenen und einzeln
käuflichen Bändchen nach Bedarf in sorgfältiger Auswahl
beigegeben.
❀
Über die bisher erschienen Bändchen vergleiche den Anhang.
 Wissenschaft und Bildung
Einzeldarstellungen aus allen Gebieten des Wissens
88

Kulturgeschichte
der Deutschen
im Mittelalter
Von

Prof. Dr. Georg Steinhausen

Bibliotheksdirektor in Cassel
Zweite neubearbeitete Auflage.

1916
Verlag von Quelle & Meyer in Leipzig
Alle Rechte vorbehalten.

Altenburg
Pierersche Hofbuchdruckerei
Stephan Geibel & Co.
 Inhaltsverzeichnis.
Seite
Einleitung: Kultur und Volkstum 1
Erstes Kapitel: Zusammenstoß und erste Auseinandersetzung
urdeutschen Wesens mit der Weltkultur 3
Zweites Kapitel: Erste Fortschritte deutschen Lebens im
Rahmen deutscher Eigenart unter wachsender Führung
der Herrenschicht (Ländlich-kriegerische Kultur) 21
Drittes Kapitel: Die stärkere Durchdringung deutschen Lebens
mit der antik-kirchlichen Kultur unter zunehmender
Beeinflussung durch die Romanen: Aristokratisches
Zeitalter 58
Viertes Kapitel: Ausbildung einer allgemeineren Laienkultur
volkstümlichen Charakters: Bürgerlich-demokratisches
Zeitalter 113
 Einleitung.
Kultur und Volkstum.

Eine nationale Kulturgeschichte muß vor anderen

Gesichtspunkten das Verhältnis von Kultur und Volkstum in den
Vordergrund stellen. Kaum einem Volke ist es beschieden gewesen,
eine Kultur ganz aus sich heraus zu entwickeln, am wenigsten aber
den Völkern, die auf jugendlich-barbarischer Entwicklungsstufe in
den Bann der antiken, d. h. der orientalisch-griechisch-römischen
Weltkultur gerieten. Je zäher und kräftiger das Volkstum, je
ausgeprägter seine Eigenart ist, um so schwieriger wird die
Auseinandersetzung mit einer siegreichen fremden, überragenden
Kultur. Insbesondere ist das deutsche Volk schwerer und später als
andere Völker zu einem solchen Ausgleich und damit zu einer
einheitlichen höheren Kultur gelangt. Die Entwicklung des
Volkstums, dessen Wurzeln zu einem guten Teile in den natürlichen
Bedingungen des Bodens und des Klimas liegen, dem aber weiter
durch langdauernde wirtschaftliche und soziale Verhältnisse
frühzeitlichen Lebens ein Stempel für spätere Zeiten aufgedrückt
wird, ist mit jenem natürlichen Grundstock und den in frühester Zeit
erworbenen Zügen nicht abgeschlossen. Das Volkstum wird auch
durch die äußere Geschichte, durch die äußere politische
Zusammenfassung selbst ganz verschiedener Völkerteile, durch die
Gemeinsamkeit der Geschicke ebenso wie der kulturellen
Verhältnisse in seiner Entwicklung infolge der Neubildung gewisser
Wesenszüge, der Änderung oder des Schwindens anderer
bestimmt[1]. Werden auf der einen Seite die fremden Elemente
höherer Kultur durch das Volkstum aufgenommen und verarbeitet,
und bewirkt die Verbindung beider Faktoren wie das
Sichdurchsetzen des Volkstums das Entstehen einer eigenen
nationalen Kultur, so wird wieder das Volkstum durch die
angeeignete Kultur beeinflußt. Aber der Verlauf der
Auseinandersetzung wird eben mit dem größeren oder geringeren
ursprünglichen Abstand der beiden Faktoren, mit der größeren oder
geringeren Anpassungsfähigkeit des Volkes, mit der größeren oder
geringeren Neigung, seine Eigenart zu bewahren, ein sehr
verschiedener. Er kann überaus wechselvoll und mannigfaltig, im
einzelnen vielfach besonders charakteristisch werden. Und das ist
bei der Kulturentwicklung der Deutschen der Fall. Man versteht
diese nicht, wenn man nicht als bestimmende Faktoren die nationale
Eigenart, das bodenständige Volkstum mit seinen Anlagen, Trieben
und alten Lebens- und Kulturüberlieferungen einerseits und die sich
auf alle Weise durchsetzende Weltkultur, deren Eindringen auch
wieder durch einen jener zähkonservativen Art entgegengesetzten
Trieb der Deutschen zu höherer Kultur, durch einen überaus
bezeichnenden Lerneifer gefördert wird, andererseits ansieht. Wie
das Verhältnis beider Faktoren den Gang der Kulturentwicklung und
seine größere oder geringere Schnelligkeit bestimmt, wie aber aus
diesem Verhältnis gerade auch die deutsche Kultur der einzelnen
Zeiten selbst als Ergebnis hervorgeht, das ist ein
Hauptgesichtspunkt der nachfolgenden Darstellung, die durchaus
nicht etwa einen Auszug aus meiner großen »Geschichte der
Deutschen Kultur« darstellt. Gerade die Geschichte der deutschen
Kultur im Mittelalter ist von jenem Gesichtspunkt aus besonders
merkwürdig. Am Schluß des Mittelalters war dann ein Ausgleich
beider Faktoren bis zu einem gewissen Grade erreicht. Aber mit dem
Einbruch neuer Wellen der höheren Kultur ergab sich ein neues Aus-
und Nebeneinander, bis das bewußte Streben nach höherer und
feinerer Kultivierung den Sieg davontrug und zu einer Kulturblüte
führte, die dem lange kulturell abhängigen Deutschland die Führung
im Reigen der Völker gab. Das wird uns später in einer Darstellung
der Geschichte deutscher Kultur in der Neuzeit beschäftigen.

Fußnote:
[1] Vgl. die näheren Ausführungen in meinem Aufsatz: Kultur und
Volkstum im »Archiv für Kulturgeschichte« Bd. VIII, Heft 2.
 Erstes Kapitel.
Zusammenstoß und erste Auseinandersetzung
urdeutschen Wesens mit der Weltkultur.

Das Gebiet des heutigen Deutschen Reiches ist bekanntlich

keineswegs von jeher von germanischen Menschen bewohnt
gewesen. Den Hauptteil des Westens, den Süden und den Südosten
hatten vielmehr, von den Germanen durch die Gebirge
Mitteldeutschlands geschieden, lange die Kelten inne. Den Osten
scheinen andererseits Letten und Slawen bewohnt zu haben, aber
sehr frühzeitig zurückgedrängt zu sein. Die ersten Sitze der
Germanen selbst sucht man, nicht ganz ohne Widerspruch
einzelner, in den Gebieten der westlichen Ostsee und auch der
östlichen Nordsee. Mächtige Bewegungen der Germanen,
hervorgerufen u. a. durch die Übervölkerung infolge der
Beschränktheit des Kulturlandes, sind also auf dem Boden unseres
Vaterlandes bis kurz vor Beginn unserer Zeitrechnung vor sich
gegangen, und auch später zeigt sich deutlich eine außerordentliche
kriegerische Beweglichkeit der Stämme, bis zur
Völkerwanderungszeit eine zweite, noch gewaltigere
Durcheinanderrüttelung und Wanderung einsetzt. Erst nach dieser
Zeit, deren Endergebnis im Verhältnis zu den heute deutschen
Gebieten auch ein gewaltiger Verlust im Osten durch das Vordringen
der Slawen war, beginnt eine zusammenhängende Entwicklung des
späteren »Deutschlands«.
Eine Kulturgeschichte der Deutschen kann erst mit dem Abschluß
der Völkerwanderung oder eigentlich erst mit der Herausbildung
eines »deutschen« Volkes in Staat und Kultur einsetzen. Die einst
germanischen Gebiete im Osten können überhaupt erst seit ihrer
späteren Kolonisierung in Betracht kommen; denn an die einstige
Siedelung der Ostgermanen kann die spätere deutsche
Kulturgeschichte nirgends anknüpfen. Im Westen und Süden waren
freilich die früher keltischen Gebiete bis zum Rhein und über den
Main hinaus bei dem Zusammenstoß mit den ein weiteres
Vordringen zunächst verhindernden Römern schon mehr oder
weniger lange von Germanen besetzt, die ziemlich frühe den
Niederrhein, gegen die Mitte des 1. Jahrhunderts v. Chr. auch den
Mittelrhein überschritten, später nach Böhmen und Ungarn drangen.
Vor allem ist ferner das »alte Volksland« zwischen Weser und Elbe
seit alters germanisch gewesen. Auch hier hat aber die
Völkerwanderung, die nur wenige Stämme, wie die Friesen, nicht
mitgemacht haben, störend eingegriffen. Immerhin haben wir im
ganzen in diesem Gebiet eine ununterbrochene Entwicklungslinie,
und so mögen einige kurze Bemerkungen über die germanischen
Zustände vorangeschickt werden[2]. Für unsere Betrachtung müssen
freilich gerade die kulturell fortgeschrittensten Germanen im Westen
und Südwesten, die zu dem römischen Besetzungsgebiet gehörten
oder mit ihm in kultureller Berührung standen, ausscheiden.
Einerseits handelt es sich hier nur zum Teil um germanische, viel
mehr um keltische Bevölkerung, andererseits fielen diese Gebiete in
der Völkerwanderungszeit der Zerstörung anheim, und nur ganz
mittelbar wirkten die einstigen entwickelteren Zustände des
römischen Germaniens in ihren Resten später nach.
Die Anschauungen über die Kultur der Germanen, wie sie gang
und gäbe sind, leiden meist an zwei Grundfehlern. Man mißt ihr
einerseits eine viel zu große Eigenart bei, und man unterschätzt
doch andererseits wieder die bereits erreichte Höhe. Viele angeblich
eigenartige Züge der Germanen sind den meisten Völkern auf
niedriger Stufe gemein. Der Erscheinung der großen, aus
unzureichendem Kulturland erwachsenen Wanderungen, die schon
Thukydides für die Urzeit vieler Völker bezeichnend gefunden hat,
entspricht der rücksichtslose Eroberungsgeist solcher Zeiten. Die
natürliche, aber auch durch wirtschaftliche Gründe bedingte
Raublust zeitigt Viehraub wie Frauenraub. Aus dem letzteren
entspringen überall die Fehden des beleidigten Stammes oder der
Sippe, und ebenso tritt überall bei den damals leicht vorkommenden
Totschlägen die Sippe des Getöteten als Rächerin an der Sippe des
Mörders auf (Blutrache). Aus dem kriegerisch-räuberischen Zug
ergibt sich auch die überall verbreitete Sklavenhaltung. Ursprünglich
verfielen diese erbeuteten Menschen dem Tode, was aber mit
sakralen Gedanken zusammentraf: das feindliche Leben wird den
Göttern geopfert. Menschenopfer sind auch für die Germanen nicht
zu leugnen. Menschenleben, fremde wie eigenes, spielen auf dieser
Stufe überhaupt keine Rolle, daher die Todesverachtung des
Kriegers, die Tötung von Greisen, das Aussetzen von Kindern, die
Grausamkeit des Vaters gegen ein zu strafendes Kind, die blutige
Härte des primitiven Strafrechts. Auch gewisse innere, seelische
Zeugnisse der ungebändigten Naturkraft sind Gemeingut der
primitiven Völker, so der Mangel an Selbstbeherrschung (im Trunk –
die Trinkfreude hat bei den Germanen freilich besondere
Ausgestaltung erfahren –, im Spiel), weiter die Unbeständigkeit, das
Fallen von einem Extrem ins andere, aus »wilder Bewegtheit« in
»starre Ruhe«, in die vielberufene »Trägheit«. Dem entspricht
überhaupt eine große Wankelmütigkeit, ein rasches Aufflammen im
Zorn, eine schnelle Beruhigung. Groß ist sodann die
Übereinstimmung auf religiösem Gebiet (Ahnenkult, Seelenglaube,
die Abwehr der Seelengeister durch Zaubersegen bei Krankheit
usw., die Verehrung von Bäumen, Hainen, Bergen, Gewässern, die
aus dem Seelenglauben sich ergebenden Totenbräuche, die
Gräberbeigaben, die Totenopfer und Totenschmäuse, das
Weissagungswesen [Befragen der beschworenen Seelen], die
Gewinnung der höheren Gewalten durch Opfer, die Verbindung von
Festen [Tänzen] und Schmäusen [Gelagen] mit den Opfern).
Gewisse sittliche Züge entspringen ebenfalls allgemein den
primitiveren wirtschaftlichen und sozialen Verhältnissen, so die
Gastfreundschaft. Auch die germanische Achtung vor der inneren
Überlegenheit der Frauen, die immerhin als eigenartig gelten darf, ist
doch nicht ganz ohne Parallelen bei den klassischen Völkern
(Frauen bei Homer). Andererseits ist bei vielen primitiven Völkern
und so auch bei den Germanen die Frau die Hauptträgerin der
Arbeit. Überall gilt auch die selbstherrliche Macht des Hausvaters,
der Frau und Kinder in einer bei den Germanen schon etwas
gemilderten Rechtlosigkeit unterstehen. Die Vielweiberei der
Naturvölker bestätigt Tacitus für die germanischen Vornehmen.
Deren Hauptmachtmittel, die keltisch-germanische Gefolgschaft,
findet sich in wenigstens vergleichbarer Form bei den Griechen (die
ἑταῖροι der Großen). Feste Formen haben die staatlichen
Verhältnisse bei allen Völkern auf dieser Stufe noch nicht
angenommen. Herrscher mit bestimmten Rechten sind die »Könige«
der Germanen nicht: die versammelte Masse der Freien ist
souverän. Noch äußert auch, wie überall, der Verband der Sippe, der
auch Kriegs- und Wirtschaftsverband ist und im Rechtsleben als
ausschlaggebender Faktor an Stelle des Einzelnen tritt, seine große
Bedeutung, wenn auch bei den Germanen bereits die Anfänge
locker-staatlichen Lebens diese Bedeutung einschränken. Auch für
die wirtschaftlichen Verhältnisse, die vielumstrittenen Agrarzustände
der Germanen vor allem, ließen sich, je nach dem Standpunkt, bei
anderen Völkern genug Parallelen finden. So vor allem bezüglich
des Gemeineigentums am Boden.
Die wirtschaftliche Kulturstufe der Germanen darf man im übrigen
nicht zu niedrig einschätzen. Heute ist man über eine
verhältnismäßige Höhe des Ackerbaus bei den Germanen im
ganzen einig. Sein hohes Alter ist ebenso nachweisbar wie eine sehr
frühe Seßhaftigkeit der Germanen. Die dem entgegenstehende,
nicht zu leugnende Beweglichkeit, die man als Beweis für
nomadische Zustände angesehen hat – auch wegen des
anscheinenden Übergewichts der Viehzucht –, meint man heute, wie
schon Waitz, als Begleiterscheinung jenes kriegerischen
Vorwärtsdringens auffassen zu sollen, wie man ja auch in der
Völkerwanderung wieder zu halbnomadischen Zuständen kam.
»Den jährlichen Wechsel der Feldmarken und Wohnsitze innerhalb
der Sippen eines Gaus zur Zeit Cäsars« muß man dann mit Hoops
als »einen kriegerischen Ausnahmezustand« ansehen.
Der wirtschaftlichen Kulturstufe entsprechen eine nicht mehr ganz
primitive Lebenshaltung (Woll-, Leinen- und Pelzkleidung, Block- und
Fachwerkhäuser, mannigfacher Hausrat, vor allem aus Holz) und
eine neben der hauswirtschaftlichen Erzeugung etwas entwickeltere
gewerbliche Tätigkeit (Böttcherei, Schnitzerei, bessere Töpferei,
Schmiedekunst, in Friesland Weberei). Eine gewisse künstlerische
Betätigung zeigt die altnationale Holzschnitzerei und Holzbemalung.
Dem Holzschnitzwerk sind auch manche Formen, wie der
Kerbschnitt, in der Metalltechnik nachgeahmt: vielleicht ist das dem
antiken Stilgefühl so entgegengesetzte Flecht- und
Verschlingungswerk, das später noch für den Völkerwanderungsstil
als charakteristisch gilt, gleichen Ursprungs.
Vom geistigen Leben der Germanen ist nicht allzu viel zu sagen.
In die Frühzeit darf man nicht allzu viel hineindeuten, am wenigsten
auf Grund der viel späteren Blüte des nordischen Geisteslebens mit
seiner gewaltigen Vorstellungswelt. Die Formenfülle der Sprache,
deren Klang den Römern naturgemäß höchst barbarisch und rauh
erschien, ist nichts eigenartiges. Auch die poetisch-sakrale Pflege
der Sprache in rhythmischer Form bei feierlichen Akten ist primitiven
Völkern gemeinsam (mit Tanz, d. h. feierlichem Schreiten im Kreise
verbundene sakrale Chorgesänge). Aus den Totenklagen durch
einzelne Vorsänger entwickelten sich episch-balladenartige Gesänge
einzelner. Episch eingeleitet wurden auch wohl die Zaubersprüche,
die Beschwörungen, episch gefärbt war auch die sonstige Spruch-,
vor allem die Rätseldichtung, deren uralte Übung jedenfalls auf kein
niederes geistiges Leben hindeutet. Eigenartig ist vor allem die
rhythmische Form der poetischen Rede bei den Germanen
gewesen, die Alliteration, die wir mit Sicherheit für eine sehr frühe
Zeit annehmen dürfen.
Trotz der betonten starken Gleichförmigkeit der Menschen und
ihrer Einrichtungen auf primitiven Stufen haben sich uns für die
Germanen überhaupt manche eigenartigen Züge ergeben. Auch was
über den späteren deutschen Menschen im allgemeinen gesagt
werden kann, darf vielfach schon auf Züge aus germanischer Zeit
zurückgreifen, insbesondere muß der Individualismus, wenn auch
nur eine rohe und unausgeglichene Form desselben, bereits für die
Germanen hervorgehoben werden. Für diesen individualistischen
Zug seien noch einige Belege hinzugefügt. Cäsar (D. b. g. IV, 1)
schon hat ihn besonders betont; er spricht von der »Ungebundenheit
des Lebens, da sie, von Kindheit an an keine Pflicht oder Zucht
gewöhnt, nichts gegen ihren Willen tun«. Und Tacitus (Ann. XIII, 54)
fügt, als er von zwei Häuptlingen spricht, die die Friesen »regierten«,
ironisch hinzu: »soweit Germanen überhaupt regiert werden.« Ein
andermal (Hist. IV, 76) heißt es bei ihm: »Die Germanen lassen sich
nichts befehlen noch sich regieren, sondern tun alles miteinander
rein nach ihrem Belieben.« Stärker kann der neben der Innerlichkeit
wichtigste Zug der späteren Deutschen nicht gut betont werden.
Es kamen die Zeiten der näheren Berührung mit den Römern, die
Zeiten des friedlichen Eindringens der Germanen in das Heer und
schließlich den Beamtenstaat der Römer, weiter die Ansiedelung
germanischer Stammesteile in dem entvölkerten Römerreich,
endlich der kriegerische Ansturm gegen dasselbe und dessen
Zertrümmerung während der sogenannten Völkerwanderung, deren
Anstoß aus dem fernen Osten kam. Für die Entwicklung der
späteren deutschen Kultur ist es nun von größter Bedeutung, wie
weit die Innergermanen – nur um diese handelt es sich für uns – in
diesen Jahrhunderten von der römischen, d. h. der Weltkultur bereits
beeinflußt wurden.
Hierbei ist zunächst festzustellen, daß südliche Einwirkungen
durch den Handel schon lange vorher bestanden. In der Bronzezeit
kamen die Bronze und das Gold nach Norden wie der begehrte
Bernstein in das südliche Europa auf alten, häufig durch reiche
Depotfunde feststellbaren Handelswegen, die, meist Flußläufen
folgend, über Land Europa durchquerten. Viel südliches vermittelten
auch die Kelten. Jetzt soll es sich aber um unmittelbare römische
Einflüsse handeln. Da sind einmal die Jahrhunderte schärfer als
bisher auseinanderzuhalten. Es ist ferner wohl zu beachten, ob es
sich um Übernahme rein äußerlicher, durch den Handel eingeführter
Dinge oder um innere, wirklich kulturelle Beeinflussung handelt. In
der Tat waren die Einwirkungen anfänglich recht gering, die
Beeinflussung des eigentlichen Germaniens rein äußerlich. Die
frühen Römerzüge in das Innere blieben ohne Nachwirkung.
Folgenreicher war die germanische Reisläuferei: mancher Söldner
mochte allerlei Römisches später in die Heimat bringen. Das
wichtigste bleibt aber der Handel. Der Handelsverkehr zwischen
Römern und Germanen scheint sich früh besonders auf den
Menschen-(Sklaven-)handel erstreckt zu haben; für Menschen
wurde Wein eingeführt. Natürlich haben die römischen, uns auch
durch Cäsar und Tacitus bezeugten Händler, die kühn sogar weit
nach Norden drangen – selbst in Skandinavien kannte man
römisches Geld –, bald auch andere Dinge gebracht, wie die Funde
zeigen, vor allem Metallgeräte und -gefäße, Schmucksachen,
Waffen, vielleicht auch Pfeffer.
Eben die Funde helfen uns besser über die Einzelheiten des
römischen Einfuhrgutes unterrichten als die Lehnwörter, auf die man
sich sonst stützt. Von den Funden zeigen zunächst die Münzen, daß
stärkerer römischer Handel vor Nero nur bis an die Ems drang und
nach dem eigentlichen Innergermanien erst gegen Ausgang des 2.
Jahrhunderts einsetzt. An eingeführten römischen oder
provinzialrömischen Gegenständen kommen in Betracht:
Bronzegefäße (italische [kapuanische Bronzeeimer und
Kasserollen], später gallische, besonders vom Niederrhein
stammend), Gläser, Tongefäße (Terra sigillata), daneben seltener
und meist erst später Gürtelschnallen, Fibeln u. dergl.,
möglicherweise auch Trinkhornbeschläge, die aber ebenso wie die
Hängezierate unter römischem Einfluß in Germanien selbst
hergestellt sein können. Daß derlei im germanischen Norden und
auch im inneren Deutschland begehrt wurde, bestätigt wieder, was
bereits über die nicht zu unterschätzende Lebenshaltung und
Kulturstufe der Germanen gesagt wurde. Die Gefäße scheinen
übrigens durchweg mit dem Weinhandel zusammenzuhängen. Was
wir sonst von Entlehnungen wissen, betrifft die römischen
Besetzungs- und Grenzgebiete an Rhein und Donau, von denen sie
erst viel später durch Vermittlung des fränkischen Staates und der
Kirche ins Innere drangen. Dahin gehören die römisch gearteten
Bezeichnungen der Wochentage (zum Teil nur Übersetzungen) und
die Übernahme der römischen Zeitrechnung; die Germanen zählten
sonst nach Nächten (noch im 16. Jahrhundert Briefschlußformel:
Hiermit viel guter Nächt) und hatten das Mondjahr. Dahin gehört
alles, was mit der Weinkultur, mit der feineren Obst- und
Gemüsekultur, der Kochkunst, auch schon mit dem Steinbau
zusammenhängt; dahin die Reihe der Münz- und
Gewichtsbezeichnungen u. a. In jenen Gegenden wird aber infolge
der innigen Berührung auch eine starke, freilich nicht zu
übertreibende und im eigentlichen Sinne wesentlich nur für die
westlichsten, wirklichen Grenzgebiete (Moselland) geltende[3]
Romanisierung des ganzen Lebens bis zu einem gewissen Grade
behauptet werden können, und das Fundmaterial soll sogar zeigen,
daß gerade die wenigen germanischen Stämme dieser Gebiete
derselben viel geringeren Widerstand entgegensetzten als die
keltischen. Wie der Saterdag (Saturntag, engl. saturday) nur in
niederrheinischen Landen vorkommt, so ist dieses den Römern
(Plinius) am besten bekannte Gebiet auch sonst eine Hauptstätte
römischen Einflusses, wie sich diesem ja schon früh die Ubier oder,
wie sie sich lieber nennen hörten, die Agrippinenser gern hingaben.
Im übrigen verfielen hier aber diesem Einfluß weniger Germanen als
Kelten; von ihnen kam durch Händler manches Römische auch zu
den Friesen und weiter zu den übrigen Küstengermanen, viel mehr
als zu den anscheinend allem Römischen und Keltischen mehr
abgeneigten Germanen am rechten Ufer des Niederrheins.
Wohl zu unterscheiden von diesem römischen Einflußbereich ist
aber Innergermanien, vor allem Nordwestdeutschland. Hier ist, wie
gesagt, in den ersten beiden Jahrhunderten n. Chr. von wirklichem
römischen Einfluß keine Rede, auch noch kaum von römischer
Handelsware. Auch Tacitus bestätigt für seine Zeit, daß nur die
Grenzgermanen allerlei Römisches annahmen und erhandelten,
nicht die Innergermanen. Der Überlandverkehr im Osten (Bernstein)
hat dorthin allerdings namentlich seit Neros Zeiten schon viel
römisches Handelsgut gebracht; ebenso mag der Seehandel den
Küstengermanen mancherlei eher vermittelt haben als denen im
Binnenlande. Vor allem zeigt dies die fehlende Beeinflussung der
heimischen Keramik. Im Laufe der Zeit aber wuchsen die Einflüsse
infolge der Beziehungen zu den römischen Provinzen und des
wachsenden Einströmens von Germanen in römische Dienste. In
spätrömischer Zeit muß z. B. die Wein- und Obstkultur schon weiter
nach Osten gedrungen sein. Man hat in Pfahlbauten in Fulda aus
dieser Zeit Trauben- und Pfirsichkerne gefunden. Später wirkten
sodann Einflüsse von der Donau her. Die die Grenzen
bedrängenden Markomannen mochten auch durch friedliche
Berührungen mannigfache Einwirkung erfahren. Am meisten
gewannen in dieser Beziehung die Goten, zumal in der von ihnen
besetzten dazischen Provinz, von der höheren Kultur, bekanntlich
auch bereits vom Christentum. Im Westen aber waren es die
neuauftretenden großen Stämme der Franken und Alemannen, die
immer heftiger die römischen Grenzlande bedrängten und sie weiter
und weiter in Besitz nahmen oder sich durch Einwanderung
festsetzten. Franken und Alemannen stellten auch das
Hauptkontingent als Söldner für das römische Heer, das zu Ende
des 4. Jahrhunderts überhaupt überwiegend germanisch war.
Trotzdem nun die Alemannen als die typischen Barbaren, als rohe,
wilde Zerstörer galten, zeigen sie ebenso wie die Franken bald
erheblichere römische Einwirkungen. Beide wurden nun zu
wirklichen Bauernvölkern; die Lebenshaltung nahm mancherlei
Römisches an. Die Alemannen hatten nach Ammianus Marcellinus
um die Mitte des 4. Jahrhunderts nach römischer Art erbaute Dörfer,
d. h. sie wohnten z. T. schon in Steinhäusern oder verwendeten bei
ihren Häusern wenigstens schon stärker die Steine. Ganz anders
steht es aber noch immer bei den Stämmen des Nordwestens, den
Friesen und den jetzt ältere Stämme vereinigenden Sachsen. Immer
bleibt es zwar bei dem erwähnten römischen Einfuhrgut: sonst ist
aber weder von höherer Landwirtschaft, Obst- und Weinkultur,
Steinbau noch von sonstigen vorgeschritteneren
Kulturerscheinungen wichtiger Art die Rede. Indessen ist auch für
die zum Teil stärker romanisierten Westgermanen, insbesondere die
Franken, festzuhalten, daß die germanische Eigenart trotz allen
äußerlichen Entlehnungen durchaus bewahrt wird. Das zeigt im
Kunsthandwerk vor allem jener sogenannte »Völkerwanderungsstil«,
auf den hier nicht näher eingegangen sei.
Das zeigt aber auch Art und Wesen der Franken selbst dann
noch, als sie nach den Stürmen der Völkerwanderung ihr großes
Reich gegründet hatten. Es sollte von all den Reichsgründungen auf
römischem Boden – von dem Sonderreich der Angelsachsen
abgesehen – allein dauernden Bestand haben und für die späteren
Deutschen eine grundlegende kulturelle und staatliche Bedeutung
gewinnen. Vom Standpunkt der Weltkultur aus gesehen bedeutete
die Zeit der Völkerwanderung, der Zertrümmerung des römischen
Reiches, zweifellos eine schwere Erschütterung. Immerhin ist von
einer Vernichtung der bisherigen Kultur nicht die Rede. Die
Zeitgenossen empfanden nichts von einem Abschneiden der
bisherigen Entwicklung. Das römische Reich ist in den Augen der
Völker nicht zugrunde gegangen: Kaisertum und Christentum