Goldstein Collingwood's Theory of Historical Knowing

Wesleyan University
Collingwood's Theory of Historical Knowing

Author(s): Leon J. Goldstein
Source: History and Theory, Vol. 9, No. 1 (1970), pp. 3-36
Published by: Blackwell Publishing for Wesleyan University
Stable URL: http://www.jstor.org/stable/2504500
Accessed: 28/02/2009 00:43
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=black.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.
Wesleyan University and Blackwell Publishing are collaborating with JSTOR to digitize, preserve and extend
access to History and Theory.
http://www.jstor.org
COLLINGWOOD'S THEORY OF HISTORICAL KNOWING
LEON J. GOLDSTEIN
In his review of The Idea of History, Maurice Mandelbaum asserts that

"Collingwood's thesis that historical knowledge consists in the historian's re-
enactment in his own mind of the thought which underlay past actions must,
I believe, inevitably lead to scepticism."' Such a belief is not simply an inde-
pendent judgment made about Collingwood's view, but rather rests upon a
conception of the nature of history2 markedly unlike the one to which
Collingwood himself subscribed. In fact, Collingwood's conclusion that the
historian re-enacts in his own mind the thought that lay behind the histor-
ically important actions he is concerned to deal with is precisely the result
of reflection designed to overcome historical skepticism. Here, indeed, is a
paradox. The identical assertion is taken by one as irredeemably skeptical
and by another as precisely the way to avoid that undesirable outcome. In
any event, it seems to me that Collingwood makes an attempt to understand
the discipline of history so as to take it seriously as a method for acquiring
knowledge of some sort. In one sense, the point of departure for this effort
is the recognition that the discipline which is history has some considerable
success to its credit, and that no account of it which makes that success un-
intelligible or rules it out as merely apparent can be taken seriously. Since
Collingwood's own early views on the subject were unambiguously skeptical,
he is clearly rejecting those views and seeking grounds for that rejection.'
1. Maurice Mandelbaum, review of R. G. Collingwood, The Idea of History, in

Journal of Phlilosophly 44 (1947), 187.
2. When I talk about history or the nature of history in this paper, I shall be dealing
with the discipline of history and not with the course of human events.
3. Thus, writers like Errol Harris ("Collingwood's Theory of History," Phlilosophlical
Quarterly 7 [1957], 35-49) and H. A. Hodges (Tie Philosoplhy of Willhelin Dilthey
[London, 1952], ch. 10), who think that Collingwood's views remained essentially un-
changed over the years are simply mistaken. Harris attempts to offer a systematic ac-
count which encompasses all of Collingwood's writings on the subject of history and
observes that Collingwood never "explicitly abandoned" the doctrine of Speculum
A'ventis (49). Hodges, too, observes that Speculum Mentis had not been "explicitly
repudiated"(338), and uses it as the touchstone for the interpretationof Collingwood's
4 LEON J. GOLDSTEIN
If, as we shall see, the method for avoiding historical skepticism character-
istic of Collingwood's mature thought involves a rejection of realism - a
theory of knowledge which he says "is based upon the grandest foundation
a philosophy can have, namely human stupidity" (Meta., 34)4_ it must
not be thought that the earlier skeptical position to which he subscribed itself
rests upon any sort of realistic foundation. One may suspect that, being dis-
satisfied with what Speculum Mentis has to say about history, he proceeded
to a serious examination of how that discipline carries out its work, and in
the course of that examination came to see how inadequate a realistic theory
of knowledge was to any proper understanding of history. But his own
initial skepticism was rooted in an altogether different sort of philosophical
commitment.
The skepticism of Speculum Mentis is rooted not in any conception of the
nature of history, but, rather, in a very restricting conception of what knowl-
edge is. Consequently, we need deal with it only in enough detail to justify
my claim that the skepticism is there. The subtitle Collingwood gave to the
book is "The Map of Knowledge," and it is clear enough from its contents
that this is intended to suggest that the work attempts to account for the
relationships that the various components of human knowledge have to one
another. Collingwood, in this earlier book, conceives of knowledge as divided
into art, religion, science, history, and philosophy, and he discusses them in
that order. But the discussion is not merely seriatim. The relationships are
taken to be dialectical, and in the course of his account of the essential
character of each - except, of course, the last - he tries to show how it
points toward the essential character of the next. For whatever reason,
science is said to point to the need for a discipline concerned to establish
particular facts, and this turns out to be history. At this stage of his thought,
history is something arrived at by way of a dialectical move of one sort or
another; it is not the discipline real historians work at. Only later could he
speak of writing down "the lessons of my last nine years' work in historical
research and reflection upon it" (Auto., 107),5 and only later will what he
has to say about the subject be important to students of the problems of
historical knowing.
Thus, the Speculum Mentis discussion of "history" proceeds with almost
no attention to the actual activities of historians, and reaches a culmination
in section five of the relevant chapter, "The Breakdown of History," which
later work. But explicitly rejected or not, it is clear enough that the Speculumn Mentis
conception of historical knowledge is simply not to be found in Collingwood's later
writings, and I do think it justifiable to claim that the burden of those writings is pre-
cisely to overcome doctrines like that of the earlier book.
4. Page references given in the text preceded by Meta. will be to Collingwood's An
Essay on Metaphysics (Oxford, 1940).
5. Page references given in the text preceded by Auto. will be to Collingwood's An
Autobiography (Oxford, 1939).
COLLINGWOOD ON HISTORICAL KNOWING 5
is a veritable subversion of the possibility of such a discipline. The opening
statement of the section sounds the keynote for the discussion which follows:
"History is the knowledge of the infinite world of facts. It is therefore itself
an infinite whole of thought: history is essentially universal history, a whole
in which the knowledge of every fact is included" (Spec., 231) .6 History is
now being judged by an external standard- a far remove from Colling-
wood's later insistence upon the radical autonomy of history - for whatever
might be said on behalf of that standard, what is said about history in the
passage just quoted could never have been learned through examination of
what historians actually do. In any event, the consequence of what is said
about history is drawn almost immediately thereafter: "History is the knowl-
edge of an infinite whole whose parts, repeating the plan of the whole in
their structure, are only known by reference to their context. But since this
context is always incomplete, we can never know a single part as it actually
is" (Spec., 231).
The function of history in the dialectical scheme of Speculum Meutis is
to provide knowledge of fact, or "objective reality independent of mind."
Such an object, supposed to be "individual, concrete, infinite, no arbitrary
abstraction or unreal fiction, but reality itself in its completeness," was sought
in vain in the previously considered spheres of knowledge. "In history we
have found it; and we have found it to be an illusion . . . . The progressive
alienation of the mind from its object is in history complete" (Spec., 238).
It is interesting to see even at this early point in the development of Colling-
wood's thought that historical skepticism is associated with a realistic ideal of
an object of knowledge independent of and unaffected by the knowing mind.
But here the matter is complicated by the conception of such an object as
being an infinite whole and the belief that the structure of each part reflects
that of the whole. The claim that one knows just a part of the whole-
given what the whole is here taken to be, no historian could ever claim more
than that -must always be mistaken, for such a part is an abstraction,
hence a distortion.
A critical examination of these bases for historical skepticism would, no
doubt, be very interesting, but I shall confine myself to showing that so far
as Speculum Mentis goes, "History as a form of knowledge cannot exist"
(Spec., 238). This is hardly a satisfactory conclusion; surely not for one who
was himself an historian. SpeculumnMentis takes the problem to be one of
how we know an infinite whole:
Every part implies the whole, and the whole is presupposedby every part. No
part can, therefore, be known first. No process of thought with respect to such
a whole is possible. We cannot come to know it. We must have known it all from
6. Page references given in the text preceded by Spec. will be to Collingwood's Spe-
culum Mentis (Oxford, 1924).
6 LEON J. GOLDSTEIN
the beginning,have known it as a whole before we began to learn any given part;
and once the whole is known any given part must be known too, and therefore
cannot be learnt. (Spec., 239 f.; italics added)
With this begins "The Transition from History to Philosophy," as the next
section is called - for the problem of knowledge must be solved even if his-
tory cannot contribute to the solution. But however the problem is taken in
Speculumn Mentis, no historian - not excepting Collingwood himself - could
leave the matter there. History is, after all, a discipline with some recognized
success to its credit, and only a most dogged commitment to the doctrines
of Speculum Mentis, or systems like it, could lead to denying the claim that
there is historical knowledge. It seems to me that the burden of Colling-
wood's reflections on history over the course of the years following the
publication of Speculum Mentis is precisely the overcoming of the skepticism
to which that work leads. And the solution is to be found not in a dialectical
move which leaves history behind, but, rather, in history itself.
II
The strange impact which The Idea of History makes upon readers, even
philosophical readers, is in large measure owing to the fact that typically
they are not sensitive to what Collingwood is actually doing. Errol Harris
is right to say that "Collingwood believes that it is impossible to determine
whether or not a proposition is true without knowing what question it is
meant to answer,"7 and it is ironic how frequently Collingwood has himself
been the victim of critics or interpreters who have not bothered to determine
what questions he was trying to answer. Thus, I think it appropriate to assert
at once that there are two possible things that The Idea of History is not:
it is not an attempt to impose an antecedent, say philosophical, conception
of knowledge upon history, though anyone who sees the work as simply an
expression of idealism or some such doctrine presumably thinks it is;8 nor
is it an attempt to characterize the course of human events - of the past
as past - though it must be admitted that there are passages which do that
7. Errol E. Harris, "Collingwood on Eternal Problems," Philosoplvical Quarterly 1

(1951), 231.
8. That Collingwood did not consider himself an idealist is clear from the following:
anyone opposing the 'realists' was automatically classified as an 'idealist,' which
meant a belated survivor of Green's school. There was no ready-made class into which
you could put a philosopher who, after a thorough training in 'realism,' had revolted
against it and arrived at conclusions of his own quite unlike the school of Green had
taught. So, in spite of occasional remonstrances, that was how I found myself classi-
fied . . ." (Auto., 56).
(cf. Idea, 119).9 Virtually fromnthe time Speculum Mentis was published,
and as longl thereafter as Collingwood thought systematically about history,
he was concerned to answer the question: what is the nature of the discipline
of history that it should be capable of yielding up historical truth? Or, how
can the historian, forever caught in the present, speak with assurance about
events which, must always be beyond his experience? Being beyond his ex-
perience is the case hinprinciple, not simply a contingent circumstance owing
to the absence of time machines and the like. Anyone who suggests that
thnefull meaning of an historical proposition (qua historical) is what would
be experienced by one who might actually be present while the event is or
was taking place, simply betrays the fact that he has not the slightest appre-
ciation of the central problem of historical knowinag.11 To my mind, the
culmination of this long inquiry is to be found in three much-criticized sec-
tions of the Epilegomena to The Idea of ilistory: "The Historical fmagination";
"Historical Evidence"; "History as the Re-enactment of Past Experience."
What I shall now try to do is show how some of their most striking doctrines
are intelligible only in the light of Collingwood's attempt to understand
what history is."1
The crucial point of Collingwoocl's conception is the recognition that the
commonplace dual reference of the word "history" is neither a mere lin-
,uistic accident nor the occasion for possible punning. The word is used
both in reference to what is supposed to have happenedI in the course of
human existence and experience and to the written accounts of these. Even
with respect to this second sense there is room for distinctions to be made;
at least we find both wider and narrower senses of "history" as referring to
such accounts. Some distinguish between annals, a mere seriatim recording
-by someone living at the time of what is supposed to lave happened during
some period, and history taken to be narrative or explanatory, or in some
other possible way a cut above mere annals. On this view, those masterpieces
of Greek literature produced by Hferodotus and Thucydides are certainly
to be counted as history. Collingwood, however, prefers to restrict the usage
9. Page references given in the text preceded by Idea will be to Collingwood's The
Idea of History (Oxford, 1946).
10. What I know through direct observation or the report of a witness, I do not
know in the historical way; cf. Colliagwood's remarks on the Greek historians in Idea,
26 f. Likewise, what I know about the past because the truth of its having happened
is vouchsafed by a revelation I deem to be inerrant, I do not know in the historical way.
11. I know that Collingwood som etimes took extreme positions, not all of which he
actually required. And I know that some of the views he expressed would have to be
rejected quite apart from the effect such reaction would have on other parts of his
doctrine. But- since my purpose in writing this paper would not be served by examining
these, I do not intend to do so. Interested readers will find some discussion of such
matters in Alan Donagan, The Later Philosophy of R. G. Collingivood (Oxford, 1962),
chs. 8 and 9.
8 LEON J. GOLDSTEIN
of the term even more. The famous Greek historians are, to be sure, dis-
cussed in The Idea of History, but they are not admitted as full historians
whose works exemplify fully the "generic essence" of history. Collingwood
discusses them as having contributed to the development of the idea of
history - clearly, though not explicitly, understood as a "scale of forms'"12
but once that idea is fully developed, they do not measure up well against
it.
Such a view of Herodotus, and especially of Thucydides, will surely be
unacceptable to any number of readers; it is not necessary to emphasize
here the great esteem with which both have been and still are regarded. But
that Collingwood should have this attitude toward their work is quite im-
portant for us here, for it emphasizes the point of his inquiry. It is not
that he is unimpressed with the narrative power or psychological insight of
the Greek historians. But he is struck by their limitations as historians:
"Their method tied them on a tether whose length was the length of living
memory: the only source they could criticize was an eyewitness with whom
they could converse face to face . . . as soon as Greek historical writing
tries to go beyond its tether, it becomes a far weaker and more precarious
thing" (Idea, 26); "the Greek historian's method precludes him from
choosing his subject . . . The only thing he can write about is the events
.
which have happened within living memory to people with whom he can
have personal contact" (Idea, 26 f.); "Thucydides' work is a %xTrfica atbe,
that of Herodotus was written to rescue glorious deeds from the oblivion
of time, precisely because when their generation was dead and gone the
work could never be done again. The rewriting of their histories, or their
incorporation into the history of a longer period, would have seemed to
them an absurdity" (Idea, 27). Some might wish to say that any criteria
according to which Herodotus and Thucydides were not accorded full honors
as historians, and which denied that their writings realized almost to the
utmost the generic essence of history, are clearly defective and ought to
be rejected out of hand. We shall not bother about that. What is to the
point is the nature of Collingwood's judgment. It is based upon recognition
that certain limitations were inherent in the fundamental methodology of
Greek historical scholarship. There were some things it simply could not
do - it is not simply an accident that their subjects were limited in the way
they were. Nor were they able to make use of what has since become known
as historical evidence. In large measure, the purpose of the first four parts
12. A term Collingwood uses in An Essay on Philosophical Method (Oxford, 1933).

Students of that work will recognize here the relevance of his view of the overlap of
philosophical concepts and his conception of the scale of forms. See particularly his
remarks on the paradox of saying both that the lower forms of a genus are members
of it yet exemplify its generic essence hardly at all (62).
of The Idea of History is to present an account of the gradual overcoming

of those limitations and the emergence in modern times of the true idea of
history.13
Some might suggest that the only way in which to answer the question,
what is history? without begging the question is to survey the writings of
historians, and to abstract from these those features which seem common
to them. Collingwood's procedure takes for granted that the discipline of
history did not arise full blown early in its own history; he contends rather
that the emergence of the fullest expression of its generic essence is the
outcome of centuries of effort. But it would not have been possible to trace
that development simply by means of careful reading of historical works.
The very possibility of the first four parts of The Idea of History having
the form they have rests upon Collingwood's already having come to some
definite conclusions about what the generic essence of history is. History
turns out to be a discipline that can do certain things, ask certain questions,
apply certain techniques. The judgment on the Greeks is based upon
Collingwood's view that they are incapable of asking those questions or
applying those techniques.14 In all this, we have the judging of historical
writing not in terms of some theory of knowledge, but, rather, in terms
of what history has become.
III
I amntrying in the present discussion to provide a context in terms of which

certain views of The Idea of History which have perplexed Collingwood's
critics would be rendered intelligible - and even seem essentially correct. And
while I have been speaking about his concentration upon the actual practice
of historical investigation, I do think it relevant to look back a bit upon
a more specifically philosophical commitment. I refer, of course, to his utter
rejection of realism in epistemology. It may be that very early in his life
he actually subscribed to realistic principles (Auto., 25),15 but he soon gave
13. Cf. Arnaldo Momigliano, "The Place of Herodotus in the History of Historiog-
raphy," in his Studies in Historiographly (Harper Torchbook ed.: New York and Evans-
ton, 1966), 127-142.
14. If Collingwood is right about the Greeks, their views would surely lead to some
sort of historical skepticism. At most, all we could know is what their historians have
recorded; we could not widen the scope of our knowledge nor would we have the means
whereby to subject their own writings to critical test. As Collingwood himself would
put it, their writings could only be treated as authorities, never as evidence for the
exercise of historical imagination. Cf. his discussion of such matters in the Epilegomena
to The Idea of History.
15. This and other relevant passages of the Autobiography are discussed in Donagan,
1 f.
10 LEON J. GOLDSTEIN
that up. By his own account, this rejection is for the most part based upon
the realization that realism is totally incompatible with what is done in
historical research (Auto., 28). I believe that realism is indeed entirely
inapplicable to history; but Collingwood's own retrospective account of the
course of his own rejection of that doctrine may perhaps be questioned.
1His Speculum Mentis is by no means a realistic work,16 yet, as we have
seen, its account of historical knowledge is clearly quite skeptical.'7 It may
well be that the order of development is the reverse of that recounted in
the Autobiography, and that his already discrediting realism made it easier
for him to discern the characteristic features of historical knowing.
In any event, it would be useful to have some idea of what the doctrine
is that Collingwood is rejecting. This is obviously not the place to present
a detailed account of the arguments for and against the various kinds of
realism in which the history of philosophy abounds. Even an attempt to
present and assess all that Collingwood himself says about realism would
take a good deal of effort. His writings are full of all manner of attacks
upon it, from comparatively well-reasoned analyses to passing, merely
polemical jibes. It will be enough here merely to state what he takes it to
be; with that done it will be easy enough to discover even in his compar-
atively early reflections on history that it is his growing anti-realism as
much as anything that puts him on the road to his mature theory of histor-
ical knowing. Or, perhaps better, it is his anti-realism which opens his
mind to an appreciation of what historical knowing is, once he has deter-
mined to devote himself to the actual study of it. For this end, some char-
acteristic observations from the Autobiography will serve as well as any.
In the relevant passage, Collingwood talks about the views of the Oxford
realists, to which he was exposed as a student, as being that knowing is
the simple apprehension of a given reality. In his view, this was essentially
16. Indeed, it already contains a version of Collingwood's anti-realistic view of know l-

edge as involving questions and answers (Spec., 76 ff.), a conception discussed in such
subsequent works as An Autobiography, Essay on Metaphysics, and The Idea of His-
tory.
17. Yet without being realistic it does contain a realistic view of history. That Col-
lingwood's earliest views of history were realistic is also evidenced by the following,
which was published in 1916, eight years before the appearance of Speculum Mentis:
"History must be regarded not as a mechanical process, nor yet as a gradual accumu-
lation of truths, but simply as objectivity; as the real fact of which we are conscious.
History is that which actually exists; fact, as something independent of my own or your
knowledge of it" (Religion and Philosophy [London, 19161, 49). (I have not seen this
book, but the passage is quoted in William M. Johnston, The Formative Years of R. G.
Collingwood [The Hague, 1967], 49; the italics are in Johnston's text.) For all that, the
actual argument for historical skepticism in Speculum Mcntis is, as we have seen, based
on an ideal of knowledge. It is because history is an infinite world of fact which can-
not be truly known piecemeal that it cannot be known at all. That it is an objective.
real world independent of our minds does not seem to enter into the argument.
the same doctrine held by realists elsewhere, whose views in general he sums
up as follows:
The condition of a knowing mind is not indeed a passive condition, for it is
-activelyengaged in knowing; but a "simple"condition, one in which there are
no complexitiesor diversities,nothing except iust the knowing. T1lheygranted that
a man who wanted to know something might have to work, in ways that migliht
in a position" from which it could
be very complicated, in order to "put hirms-eIf
be "apprehended";but once the position had been attaided there was nothing for
him to do but "apprehend" it, or perhaps fail to "apprehend" it. (Auto., 25 f.)
Knowledge, on the realistic view, turns out to be an immediate or intuitive
grasping by a subject of an object present to its consciousness. It is this
idea of immediate or direct apprehension, which he takes to be a charac-
teristic mark of realistic theories of knowledge, that Collingwood thinks is
too simplistic to be taken seriously. tlow ironic that his critics, in their
bafeflemnt over the sense of his well-known aphorism "When he [the his-
torian] knows what happened, lie already knows why it happened" (Idea,
2114), should seek to saddle such a view on him!"8The objects of historical
study, at any rate, are never given to immediate apprehension.
IV
For all its unfinished character, The Idea of History is the mature statement
of Collingwood's reflections on threnature and problems of historical know-
ing, as we shall soon see. But there is evidence of the direction his thought
was taking in his earlier writings, now happily brought tog-ether from the
scattered periodicals of their original publication within the covers of one
volume.19 In a paper called "The Nature and Aims of a Philosophy of
History,9 which goes back to the period of Speculant Mtleuatis,but appai-ently
does not reflect the holistic conception of knowledge we found in that work,
Collingwood remarks that "the battle of Hastings is a label for something
which, no doubt, did happen in that year [1066]; but no one knows, no
one ever has known, and no one ever will know what exactly it was that
happened" (Essays, 43). But instead of expressing the skepticism we might
have expected from the author of Speculum Mentis, what immediately fol-
18. I do not say that if you understand Collingwood's statement you will accept it;
cf. Donagan, 201. But I do say that anyone who cites it in support of the claim that
Co]lingwood believed that one could grasp an event and its explanation in one imme-
diate intuition does not understand it. To understand it means to see it as Collingwvood-
ian, to determine what someone holding Collingwood's views must mean by it, not to
analyze it as an ordinary English sentence having no context.
19. R. G. Collingwood, Essays in the Philosophy of History, ed. with an introduction
by William Debbins (Austin, Texas, 1965); page references to this volume will be given
in the text preceded by Essays.
lows is an explicit denial that what he said is skeptical; "for scepticism

implies that no opinion is preferable to any other; and it is certainly possible
to choose between different historical views . . . without implying that one
knows all there is to be known about it oneself."
In Speculum Mentis, historical thought is a dialectically grounded moment
in the course of an account of the nature of knowledge. In "The Nature and
Aims of a Philosophy of History," history and its philosophy are the objects
of immediate interest. In the first, history is not attended to for its own sake,
and, measured against a criterion of knowledge which finds it wanting, it is
abandoned. In the second, however, history is too important and taken too
seriously to be rejected merely because on some possible criterion - ex-
ternal to itself - skepticism seems to be called for. It is of some interest to
see just how this skepticism is rejected. Soon after the remarks quoted in the
previous paragraph, Collingwood says:
Ideally, historical thought is the apprehensionof a world of fact. Actually it is
the presentationof thought to itself of a world of half-ascertainedfact: a world
in which truth and error are at any given moment inextricablyconfused together.
Thus the actual object of actual historical thinking is an object which is not
"given"but perpetuallyin process of being given. (Essays, 44)
This distinction between the ideal and the real would not have been possible
in Collingwood's later writings, for as he became increasingly preoccupied
with what the discipline of history is, he became increasingly unwilling to
allow that its autonomy could be limited by standards outside of itself. Here,
however, he seems still willing to allow that the common-sense ideal of
history as an attempt to describe a past as it was, a reality external to the
thought of the historian, is acceptable. But while a goal to be striven for, it
is never achieved. The object of investigation is never given fully and finally,
but is, rather, "perpetually in process of being given." This is what history
is, and accepting history as a legitimate way of knowing, the Collingwood of
"The Nature and Aims of a Philosophy of History" accepts it with all of its
shortcomings. It is interesting to note that even here, with his explicit accep-
tance of the common-sense ideal of historical thought as an "apprehension
of a world of fact" - an ideal which is implicitly realistic since such a
world of fact is presumably there, unaffected by the consciousness of the
knower -he is already moving toward his autonomous and non-realistic
conception of history as a discipline whose object is the "presentation by
thought to itself of a world."'2
20. Collingwood's conception of the autonomy of history is very important for our
understandingof his views and, in particular, is central to his attempt to overcome his-
torical skepticism. Thus, it is worth making certain that we understand what it is and
what it is not. As will emerge in the course of our discussion, Collingwood's position
is that the discipline of history is autonomous, in that it is not dependent on things and
C()LLINGWOOD ON HISTORICAL KNOWING 13
Later in the essay we have been considering, Collingwood compares
history and perception in such a way as to anticipate what he will say in
The Idea of History about the historical imagination. I shall not develop
that comparison fully since his views on perception as such are not germane
to what interests us here. For him, perception involves a large component of
interpretation -something not at all surprising in a thinker who was in-
creasingly rejecting realism on the one hand, and who so much emphasized
the autonomy of history and the central importance of the historian's imag-
ination in the reconstruction of the past on the other. In any event, it does
seem to me that here we see in an early form what is to become his strong
emphasis on the historian's imagination, but unlike the later doctrine, which
is concerned with the reconstruction of past events, this earlier form is limited
to the interpretation of evidence.
It may seem odd to say about history that it "in its fundamental and ele-
mentary form is perception" or that it "is perception raised to its highest
power" (Essays, 49). This is because "Perception appears to the perceiver
as immediate" (Essays, 49), and some philosophers have simply adopted
the standpoint of the naive perceiver. For Collingwood, however, reflection
makes clear that in perception we have to do with two elements, sensation
and thought, and that sensation is only an abstraction, "the limiting case in
which we are supposed to receive unreflectively a pure datum. In actual
experience we never get such a pure datum: whatever we call a datum is in
point of fact already interpreted by thought" (Essays, 50). In view of this,
what at first sight may have seemed to be the strange comparison of two
fundamentally disparate modes of mental experience markedly changes its
character. "The only difference between what we ordinarily call perception
and what we ordinarily call historical thinking is that the interpretative work
which in the former is implicit and only revealed by reflective analysis is in
the latter explicit and impossible to overlook" (Essays, 50). The parallel
criteria outside of itself in order for its work to proceed. It is not in the least what
Patrick Gardiner thinks it is. His discussion of it is presented in the course of what is
supposed to be an account of views to the effect that history is sui genesis, a view
which he grounds in belief that history deals with its own peculiar subject-matterwhich,
he says, Collingwood treats as "a self-contained world that must accordingly be inter-
preted by methods bearing little or no relation to those used in other branches of knowl-
edge." Gardiner makes a number of mistakes here. Collingwood's view is that the dis-
cipline of history is autonomous, not that some peculiar historical entities -the "world
of experience"- are autonomous with respect to other entities. If Gardiner were correct
in his reading of Collingwood, then the discipline of history would not be autonomous
at all; instead, we would have the subordination of history to the "dictates" of some-
thing external, namely to the separate spheres of existence Gardiner says Collingwood
believed in. Finally, far from claiming special methods of its own -"esoteric methods"
as Gardiner calls them - we shall see that for Collingwood the method of history is
thinking or re-thinking, surely not a method confined to history alone. See Patrick
Gardiner, The Nature of Historical Explanation (Oxford, 1952), 28-34, 49.
1-4 LEON J. GOLDSTEIN
between these two modes of experience is made even more explicit in the
words which soon follow: "in all perception we are making a judgment,
trying to answer the question what it is that we perceive, and all history is
simply a more intense and sustained attempt to answer the same question."
The rationale for this reflects his rejection of historical realism, his refusal
to take seriously the general view that history is concerned with a real past,
complete and simply waiting to be discovered. "The past events which the
historian brings to light are only revealed by his thought in its attempt to
understand the world present to his senses: a past event which has left no
trace on his perceptible world is to him unknowable" (Essays, 50).
In the account of the previous paragraph we see easily enough the direc-
tion Collingwood's thought is taking in his attempt to overcome historical
skepticism. The point of the claim that historical thinking is like perception
is that it is an affair of the present; that it is concerned with what is now
given to the historian's senses. That which is given to the historian's senses
is not a past event, but rather historical evidence."' We shall see in a moment
just what Collingwood has to say in this comparatively early essay about
historical evidence and its interpretation, but it is worth pointing out that
the character the discussion takes makes it evident that Collingwood has not
yet reached the culmination of his theory of historical knowing, with its
central attention to the role of imagination, not in the interpretation of ev-
idence, but in the reconstruction of events. I venture to suggest, and hope that
it will become more evident in the sequel, that until Collingwood came to
recognize the role that the historian's imagination plays in the discipline of
history, the historical event, with its seeming pointing to a forever unap-
proachable past-in-itself, may have been somewhat embarrassing for him. Be
that as it may, the discussion which follows, informed as it is by a non-
realistic conception of perception, deals with the historian's interpretation
of what he qua historian perceives, namely historical evidence.
Historical thinking, according to the essay we have been considering, is
critical thinking about evidence: "We have not only to read, but to criticism"
(Essays, 51). Perceptual judgment rests on past experience, "and we per-
21. The problem of the past in the context of concern with historical knowing is how
we who are always caught in the present, having no direct access to the past, can claim
to have knowledge now of what happened then. The whole point of the discipline of
history is to enable us to make warranted assertions about a past we can never know
by acquaintance. It is failure to attend to this - and, more particularly, failure to rec-
ognize how aware of this Collingwood was -that surely lies behind Gardiner's view
that the problem for Collingwood was to have a past that was present. If the past could
be known by acquaintance, then the discipline of history and its concern with historical
evidence would be of no use to anyone. One may presume that the same tendency that
leads Gardiner to foist upon Collingwood belief in peculiar historical entities (see the
previous note) is at work here too. See ibid., 39 f.
ceive more and more accurately according as we become more and more
able to compare the present experience with relevant experience in the past"
(Essays, 50 f.). Similarly, historical thinking requires experience, initially in
the form of relevant training of the sort that will make it possible for the
practitioner of the discipline to make critical assessments of historical ev-
idence. Progress in historical study involved the refusal of critical historians
to believe something simply because it is written in some old document.
Rather, by dint of application of the historical disciplines which deal with
evidence, historians have learned how to question sources, rejecting asser-
tions which cannot be true and extracting from them all manner of un-
dreamed-of truth. "No competent historian who reflects on the progress of his
own thought can overlook the way in which progress has created22 masses
of evidence bearing on questions concerning which there was once no ev-
idence whatever" (Essays, 53). Several years later,23 Collingwood could
still say:
For historical thinking means nothing else than interpretingall tho available ev-
idence with the maximum degree of critical skill. It does not mean discovering
what really happened,if "what really happened"is anything other than "what the
evidence indicates."If there once happened an event concerning which no shred
of evidence now survives, that event is no part of any historian'suniverse it is
no historian'sbusiness to discover it; it is no gap in any historian'sknowledge
that he does not know it. (Essays, 99)
From the standpoint of Speculum Mentis, this last passage is simply im-
possible,24 for belief in the existence of past events concerning which his-
torians have no knowledge is precisely what leads Collingwood to historical
skepticism in that book. We have seen that the problem his thinking about
history is intended to resolve is precisely that of skepticism, and we have
seen that the means whereby that resolution is accomplished is the attending
to what historians do. And that means, at least during a large part of the
period between Speculum Mentis and The Idea of History, that history seems
at least implicitly to be understood not as that discipline which concerns
itself about the formulation of warrantable assertions about past events, but,
rather, as the discipline whose subject matter is historical evidence - those
data which exist now, in the present, which historians have to understand
and interpret. Thus, he can say: "The past is simply non-existent; and every
historian feels this in his dealings with it" (Essays, 101). Most interesting is
22. At this point, Collingwood adds the following footnote: "Created,not discovered,
because evidence is not evidence until it makes something evident" (his italics).
23. In a paper called "The Limits of Historical Knowledge," published in 1928.
24. And underscores how mistaken are those who think that Collingwood's views
remained essentially unchanged and that his later writings are to be understood in light
of that earlier work.
the way in which his view of the very basis of historical skepticism has
changed. He does say at one point that historical skepticism has a useful
cognitive function - not to call into question the validity of historical knowl-
edge, but rather to point to its limits (Essays, 91) - but that is not the
point I find so striking. What is of interest here is that historical skepticism
is laid at the door of historical realism itself. "The discovery that the past
as such is unknowable is the skepticism which is the permanent and neces-
sary counterpart of the plain man's realism" (Essays, 100). The point here
is that skepticism is the consequence of any conception of history which
makes the focus of attention not what historians do in the present, but what
really happened in the past, a conception which takes as equally factual,
hence historical, the question what Caesar had for breakfast some given
morning and the question whether he intended to become king of Rome
(Essays, 101). Curiously, Collingwood has not forgotten that earlier, Spe-
culum Mentis basis for skepticism, the view that knowledge is and can only
be an indivisible whole. But now he saddles a modification of that view on
the realists. Historical realists are taken to be those who measure success in
historical work in terms of the quantities of fact discovered. Their episte-
mology commits them to a view according to which knowledge is the grasp-
ing or apprehension of objects which exist independently of the knower. In
history this is possible only if all of the past events of human experience
still exist in some peculiar way,25 and historical knowledge is the apprehen-
sion of them. The more such facts are known, the closer do we come to the
ideal of knowing the whole (Essays, 99 ff.).
The direction Collingwood's historical thought was taking during the
period from Speculum Mentis to The Idea of History is now clear. Increas-
ingly he came to think that the way to avoid the pitfalls of skepticism was
to move away from an unknowable past-in-itself and pay rather close heed
to what historians do. This, of course, rests on the assumption that historical
skepticism is to be rejected on the ground that what historians do and have
accomplished justifies the claim that history is a legitimate way of knowing.
Historical skepticism, it would seem, must always be consequent upon stand-
points outside of history itself - say, an holistic conception of knowledge
or a realistic view of knowing - but cannot arise from consideration of
history itself, as it is practiced by historians.26 I should not wish to suggest
25. See Collingwood, "Some Perplexities About Time: With an Attempted Solution,"
Proceedings of the Aristotelian Society 26 (1925-26), 144, 146, and 150.
26. The skepticism of Charles A. Beard stems not from his reflections on historical
method, but from his commitment to goals, rooted in historical realism, with which
those methods cannot be reconciled. Note his repeated insistence that the interest of
history is in the "past as actuality" in his "Written History as an Act of Faith," in
The Philosophy of History in Our Time, ed. H. Meyerhoff (New York, 1959), 140-151.
that in Collingwood's opinion philosophical considerations are merely philo-
sophical considerations, for I think that strewn through his writings are all
manner of suggestions to the contrary. But it is nevertheless true that it was
only after he began to think about the procedures of history itself that he
was able to see his way clear of skepticism, and it is likewise true that he
came early to believe that one virtually fatal flaw of realistic epistemologies
is their incapacity to do justice to history.
I want now to turn to The Idea of History in order to present and discuss
a number of its most important -some would say most outrageous -
doctrines. As is well known, this book was published posthumously. Two-
thirds of it, the first four of its five parts, are a critical historical sketch of
the development of the concept of history - the emergence through time, as
it were, of the generic essence of the discipline of history. The final part, to
which our attention here will be limited, consists of a number of separate
essays. The actual volume which is called The Idea of History was put
together out of these parts by Sir Malcolm Knox, who explains what he did
in the "Editor's Preface." All this results in a work the parts of which come
from different times of the author's life - though none from what might be
called his early period -and which did not have the advantage of the
author's final editorial scrutiny. It is very easy to single out passages from
The Idea of History for criticism on one basis or another. But for all of that,
and for all its incompleteness, the general drift of the book is toward a
conception of historical knowing which is essentially sound. Let us set the
stage for presenting the main elements of that conception by presenting an
extract from the book (Idea, 132 f.):
The conception of history as dealing with fact may seem harmless enough, but
what is a fact? According to the positivistic theory of knowledge, a fact is some-
thing immediatelygiven in perception.When it is said that science consists first in
ascertainingfacts and then in discerning laws, the facts, here, are facts directly
observed by the scientists. . . . If anyone doubts the fact he can repeat the ex-
periment . . . ; and consequently,for the scientist, the question whether the facts
really are what they are said to be is never a vital question, because he can al-
ways reproducethe facts under his own eyes. In science, then, the facts are em-
pirical facts, facts perceived as they occur.
In history, the word "fact" bears a very different meaning. The fact that in the
second century the legions began to be recruited wholly outside of Italy is not
immediately given. It is arrived at inferentiallyby a process of interpretingdata
accordingto a complicatedsystem of rules and assumptions.A theory of historical
knowledge would discover what these rules and assumptionsare, and would ask
how far they are necessary and legitimate. All this was entirely neglected by the
positivistic historians, who thus never asked themselves the difficult question:
How is historical knowledge possible? How and under what conditions can the
historian know facts which, being now gone beyond recall or repetition, cannot
be for him objects or perception?They were precludedfrom asking this question
by their false analogy between scientific facts and historical facts. Owing to such
a false analogy, they thought such a question could need no answer. But, owing
to the same false analogy, they were all the time misconceiving the nature of
historical facts, and consequently distorting the actual work of historical re-
search
The distinction between the facts of history and those of natural science
certainly helps make clear the problem, even though I rather doubt that the
distinction as presented here is one that Collingwood really wished generally
to make. For the quoted passage would make it appear that Collingwood
was willing to concede that realism was an adequate epistemology for the
natural sciences, yet that is surely far from the case. Collingwood wants to
say, here, that positivist historians assimilate historical facts to natural facts
and so never allow themselves to raise the question as to how the former
can be known. His realistic critics might urge against him that his own
tendency is to do the reverse, to assimilate natural fact to historical fact and,
consequently, lose the independent existence of the real world. Yet even if
what Collingwood says here is not fully consistent with what he says about
realism elsewhere, his way of putting it here is perfectly serviceable. Make
the strongest possible case for realistic theories of knowledge, and accept
everything realists tell us about our knowledge of the world around us and
the starting points of natural scientific inquiry; yet when it comes to the facts
of history-which in the passage quoted clearly means historical events of
the past, not historical evidence - realism has nothing to say. The assimila-
tion of historical facts to natural facts amounts to asserting that the inacces-
sibility of past events is merely accidental or fortuitous. We can look at
what is happening around us, but not at what happened long ago. Some
writers seem to think that the difference is merely fortuitous, and the full
meaning of a description of a past event is precisely what would appear to
an observer, or ideal observer, witnessing the event.27 But a God's-eye view
of the course of human events would not be an historical view of it, nor
would an account certified as true by God, i.e., a work of divine revelation,
be an historical account. An historical account is one which is the outcome
of the application of the methods and techniques of historical inquiry, and
these do not include the reports of ideal observers. The reports of such
observers and the reconstructions of historians have clearly different epistemic
27. Cf. Richard M. Gale, "Dewey and the Problem of the Alleged Futurity of Yes-
terday," Philosophy and Phenonienological Research 22 (1961-62), 501-511, and my
"The 'Alleged' Futurity of Yesterday," ibid. 24 (1963-64), 417-420.
statuses. By "the facts of history" Collingwood would understand facts estab-

lished by historical research, not the actions to which human beings have
been party over the course of their existence on earth. From which it clearly
follows that historical facts cannot be facts in the realistic or positivistic
sense.28
This is a very important issue, and no small part of the critical reception
of Collingwood's philosophy of history may be traced to a tendency on the
part of many not to attend to the epistemic differences between present and
past descriptions. What I mean may become clearer if we consider briefly the
views of a writer who, far from agreeing with Collingwood, is quite critical
of him: A. M. MacIver. In Maciver's view, "We contrast the 'generality' of
historical statements with the individuality of facts on which they are
based"29 and he goes on to show what he means in the following passage:30
A typical historical statement is "The Normans defeated the English at Hastings
in 1066." The battle itself was a vast medley of individualactions and experiences
- this man shooting this arrow, that man avoiding it or being hit by it, horses
stumbling,men feeling pain or fear or exultation-but the historical statement
takes it as a whole and selects for mention just that aspect of it which bears upon
the historian'spurpose- in this case, the fact that as a result the Duke of Nor-
mandy was able to make himself King of England. It is not the business of the
historian to "generalize"in any other sense than this.
As an account of what historians do, the above is patently false, as I shall
make clear shortly, but let us at least indicate what about it is right, what
would lead most readers to think it sound. And that is that anyone who
sought to answer such questions as "what is a battle?" or "what is an army?"
would have to speak about the sorts of thing MacIver does. In a battle
soldiers behave as they are wont to do, and it would be utterly absurd to
think that you can have an army without soldiers or a battle without shooting
and being shot -with whatever projectiles could be managed at various
stages of military technology.
But an historian's statement with respect to the Battle of Hastings is not
generalized in the way MacIver says; it is not based on the individual facts
to which he adverts. Let us be clear about the fact that we are no longer
28. Hence, pace Gardiner, Collingwood could not possibly subscribe to the view that
historical knowing was knowing by acquaintance (Gardiner, 39). Indeed, how anyone
as opposed to realism in any sphere of knowledge as Collingwood was could be thought
to believe that the acquaintance theory had any application at all is beyond my com-
prehension. His following remarks are pertinent here: "All theories of knowledge that
conceive it as a transaction or relation between a subject and an object both actually
existing, and confronting or compresent to one another, theories that take acquaintance
as the essence of knowledge, make history impossible" (Idea, 233).
29. A. M. MacIver, "The Characterof a Historical Explanation,"Aristotelian Society
Supplementary Vollume XXI (1947), 38.
30. Ibid., 39.
talking about the nature of armies and battles in general, about the "ontol-
ogy" of armies and battles. We are now concerned with the nature of his-
torical knowing, with the epistemic status of the historian's statement. On the
basis of evidence, the historian concludes that a battle took place, that the
Normans were victorious, and that William succeeded Harold as England's
king. Far from basing his assertion that a battle took place on the sort of
individual facts Maciver mentions, the historian does not even know what
they were. He does not know anything about who shot what arrow at whom
and what resulted therefrom. Except for the major participants, he does not
even know who was there, what he did and how he emerged. To be sure,
being in full possession of his senses and knowing in general what is involved
in warfare of the sort in question, he does believe that individual soldiers
performed individual acts of soldiery. What could he mean by the assertion
that a battle took place if not that? Yet his knowledge that a battle took
place is not generalized from these acts because he does not know what these
acts were. Epistemically, the basis for the claim that a battle was fought was
the evidence -the precise relation between evidence and the knowledge of
historical events need not detain us here.3'
If the Battle of Hastings were directly accessible to the historian's ob-
servation, then perhaps it would be a technique of historians to generalize
in the way MacIver says, and there would be no difference in status between
past and present descriptions. But then our present discipline of history as a
form of knowing would be entirely superfluous. It is the absence of such
access to the past that necessitates the historical form of knowing if we are
to be able to speak of the past at all. Collingwood was fully aware of this,
and it is that awareness which lies back of the views we are now to consider.
VI
It is now possible to understand the point of the main doctrines of The Idea
of History, and it is possible to understand the way in which they emerge
from Collingwood's reflection both on what historians do and how history is
to be justified as a discipline capable of yielding warranted results. This
will become clear as we examine the main ideas in the sections of the last
part of The Idea of History which were mentioned earlier in our discussion.
We shall begin with the notions of the historical imagination and the
autonomy of history which are presented in the section called "The Historical
Imagination."
To say of any discipline that purports to be about something or other
31. But cf. my "Evidence and Events in History," Philosophy of Science 29 (1962),
175-179.
that it is autonomous, that it is not dependent upon things outside of itself,

seems highly dubious at first blush. Surely, we want to reply, no science is
free to do just what it wants; rather, it is clearly dependent upon the nature
of that which it studies.32 It is evident enough by now how Collingwood
would receive such a remark. Since the objects of history "are events which
have finished happening, and conditions no longer in existence" (Idea, 233),
it is simply not possible to claim that something now in existence, namely, the
enterprise of history, depends upon it. But there is another way in which
Collingwood's critics might oppose the claim that history is independent, and
that is by appeal to what Collingwood calls "the common-sense theory" of
history, which he presents as follows (Idea, 234 f.):
Accordingto this theory, the essentialthings in history are memory and authority.
If an event or state of things is to be historically known, first of all some one
must be acquainted with it; then he must remember it; then he must state his
recollectionof it in terms intelligibleto another;and finally that other must accept
the statement as true. History is thus the believing someone else when he says
that he rememberssomething. The believer is the historian; the person believed
is called his authority.
I dare say this theory could be expressed in other ways, with this point
qualified, that modified, and some other one changed in some other way.
The central point of the theory, which would presumably survive all modifica-
tion, is that the historian may not gainsay the actual claims of his author-
ities.33 But this is totally false to the practice of historians. There is nothing
contained in any of his documents that the historian cannot doubt, or even
reject. And this, of course, means that whatever criteria the historian uses in
terms of which to reject or accept the assertions of a document, those criteria
are independent of the document itself. In addition, the use historians make
of evidence which does not make assertions - ruins, coins, artifacts -
makes clear that they have ways of determining what happened that have
nothing at all to do with believing the claims of authorities. All of these ways
and criteria are part of the present methods of historical research, and that
the historian may apply them to his evidence, may subject his evidence to
their control, is the measure of the autonomy of history.
In Collingwood's actual discussion of these points (Idea, 236), there is a
tendency to ignore an important distinction between two senses of mind-de-
32. However, should we characterize sciences, not in terms of the objects they study,
but in terms of their conceptual systems, we could argue that they determine, concep-
tually, the nature of their own subject matters. But there is nothing to be gained here
by pushing this point, and it is perfectly satisfactory to allow the character of the sub-
ject matter of natural science and that of history to continue to appear to be basically
different. Cf. my "Ontological Social Science," American Anthropologist 61 (1959),
290-298.
33. Cf. Gale, 509.
pendence. I have just spoken of the autonomy of history, but he writes of

the historian as being his own authority and of his thought as being "auton-
omous, self-authorizing, possessed of a criterion to which his so-called au-
thorities must conform and by reference to which they are criticized." In
the paragraph immediately following, he compares the historian to an artist:
"It is the artist, and not nature, that is responsible for what goes into the
picture. In the same way, no historian, not even the worst, merely copies out
his authorities .... It is he, therefore, and not his authority, that is respon-
sible for what goes in" (Idea, 236 f.). Soon after this, he talks about the
historian's "a priori imagination" (Idea, 240 ff.), saying that much of what
he takes to be true he constructs by means of his a priori imagination, and
then adding: "I am more driven to confess that there are for historical
thought no fixed points thus given: in other words, that in history, just as
there are properly speaking no authorities, so there are properly speaking
no data" (Idea, 243).34 Surely, then, those critics who see in The Idea of
History a doctrine of skepticism rooted in the author's inability to get beyond
the subjectivity of each historian seem vindicated.
In my opinion, this conclusion of the critics is mistaken. My purpose
here, however, is to point to the nature of Collingwood's mistake, for I do
not think that the line of thought about history that he developed requires
the subjectivism he sometimes seems to espouse.35 The source of the error
is failure to keep distinct two senses of mind-dependence, the idiosyncratic
and the non-idiosyncratic. Mind-dependence in the former sense refers to
dependence on the peculiarities of an individual, rather like the way in which
the blurred look of the world when I am not wearing my glasses depends
upon my myopia or someone's conception of the course of some historical
event upon his narrow-minded prejudice. In the latter sense, it is like the way
in which a good deal of human action depends upon shared meanings,
values, and institutional arrangements for its intelligibility. By speaking of it
as mind-dependent we mean that it is not merely or directly a creation of
nature, but we do not mean to suggest that what we deal with depends upon
the idiosyncrasies of particular individuals. It is a regrettable fact that not
34. For an earlier expression of subjectivism, see his 1930 paper, "The Philosophy
of History" (Essays, 137 ff.).
35. Cf. "Everyone brings his own mind to the study of history, and approaches it
from the point of view which is characteristic of himself and his generation .
The attempt to eliminate this 'subjective element' from history is always insincere
and always unsuccessful. If it succeeded, history itself would vanish." This leads Col-
lingwood to what is surely a most vicious form of subjectivism: "This does not reduce
history to something arbitrary or capricious. It remains genuine knowledge. How can
this be, if my thoughts about Julius Caesar differ from Mommsen's? Must not one of
us be wrong? No, because the object differs. My historical thought is about my own
past, not about Mommsen's past" (Essays, 138 f.).
infrequently these two senses are confused. Perhaps it is because when we
think of mind we think of an individual's mind, and we are often wary lest
we be charged with hypostatizing group minds and the like. And it has often
proved difficult in the English-speaking philosophical community to think of
thought or meaning without seeking at once to reduce the whole matter to the
thought of some thinker or the meanings of some actor. But it is certainly
possible to distinguish the content of thought and the thinking of thought,
and any number of writers have held the view that the very possibility of
such disciplines as social and intellectual history, cultural anthropology, and
sociology depends upon the feasibility of this distinction.36
What Collingwood ought to say is that history is mind-dependent in the
non-idiosyncratic sense, as can be seen from a summary of the argument
so far. The task Collingwood set himself is the vindication of history against
any forms of skepticism which would call it into question, and this he does
by attending to what history is. He does not ask, "is history possible?" but,
rather, "what is history that it is possible?"37 The answer which emerges in
the course of his work is that it is a discipline having certain procedures and
methods, which makes use of what is called historical evidence, which is
autonomous in that it is not dependent on criteria other than those it estab-
lishes, and so forth. In all these ways it is mind-dependent in the non-idio-
syncratic sense. But like anything mind-dependent in that sense, it has no
existence apart from individuals: it would be ludicrous to think of the
discipline of history without historians. Yet the discipline is not dependent
upon the peculiarities - much less the biases - of any of its practitioners,
and the mistake Collingwood makes in the passage we have been considering
is the failure to keep distinct the two senses of mind-dependence and the
failure to keep from slipping from the discipline and its characteristics to its
practitioners and theirs.
But in view of the way I have been supporting Collingwood's rejection of
historical realism, virtually daring his critics to produce a real fact-in-itself
to be the subject matter of historical research, some may wonder if I have
36. Dilthey was certainly one of those who held to the distinction, but the reader of
the collection of excerpts from his writings on history, Pattern and Meaning ineHistory,
transl. and ed. H. P. Rickman (New York, 1961), may sometimes have difficulty in
keeping this in mind owing to Rickman's decision (23) to translate Dilthey's "das
Geistige" as "mental content." This often makes it seem that Dilthey's focus was on
the idiosyncratic-psychological when actually he was concerned with social-historical
content; cf. 69, 71, 72, and 76. I cite this not to be critical of Rickman, but only as
an example of how easy it is to slip from the non-idiosyncratic kind of mind-depen-
dence to the idiosyncratic kind.
37. Dilthey, too (ibid.), takes the possibility of history for granted, but his approach
to answering this second question is, in the Kantian manner, to look for the categories
which are constitutive of the historical world.
a right to appeal to the distinction of the two senses in order to keep non-
realism from falling into subjectivism. Briefly, such wondering might proceed
in the following manner. If the real past is, indeed, not available to be
examined or studied, what, in fact, is? Clearly, historians - individual
scholars - engaged in historical research. Each historian is a subject em-
bodying a standpoint in terms of which his work is carried out. Each does his
own investigating, makes his own judgments, accepting or rejecting according
to whatever criteria he chooses to accept. Having rejected historical realism,
the critic claims, Collingwood has at the same time made it impossible for
us to have a non-subjective history; each historian, his imagination as un-
restrained as that of a novelist, may simply do as he pleases.
But while we have historians - individual subjects, to be sure - we also
have history. We have an organized structure of historical research. We have
widely-shared techniques and generally-agreed-to results. We have, in sum,
an inter-subjective discipline, and while it is carried on by individuals, they
are committed to a common enterprise having determinate features. The
claim, then, that history is autonomous need only be made with reference to
that discipline. History is autonomous in a way historians may well not be.
It is, then, the autonomous discipline of history which controls and has the
final say about historical evidence, and that discipline is not subject to the
authority of old texts. With this much said, we may now turn to the next
section of The Idea of History, "Historical Evidence."
VII
The relation of history -- the discipline - to historical evidence is one of

the pervasive themes of The Idea of History, and perhaps more than any-
thing else it is the changing character of that relation which marks the dif-
ferent stages of the evolution of the idea of history. I noted earlier that in
Collingwood's view the concept of history is not subject to precise definition
but is rather a scale of forms the generic essence of which is increasingly
realized in the course of the history of history. At first, there is virtually
complete dependence of the historian upon his sources, and, it will be recalled
that Collingwood says of the Greek historians that they were capable of
using only one kind of source, the reports of witnesses. Even when historians
could use written reports of witnesses no longer living, hence no longer
available for questioning, at first, Collingwood claims, they could either
incorporate the content of the report or reject it, but it took some time before
they arrived at techniques for dealing with it critically. As it matures - and
more perfectly expresses the generic essence of its concept history learns
to subject its sources to its own purposes, to extract from a text all manner
of information not explicitly asserted in it, to deny its sources all claims to
authority. History, itself, as it becomes increasingly autonomous, becomes
the sole authority: far from relying on the reports of others, it is history
itself which decides what is and what is not historical evidence and what
possible use this or that historical evidence may have.
According to a widely-held, yet clearly naive, view, the historian discovers
his facts by reading sources: what he may assert with warrant is limited by
the testimonies they preserve. In Collingwood's view, such testimony cannot
be historical knowledge:
What I assert is that it can never be historical knowledge, because it can never
be scientific knowledge. It is not scientific knowledge because it cannot be vindi-
cated by appeal to the grounds on which it is based. As soon as there are such
grounds, the case is no longer one of testimony. When testimony is reinforced
by evidence, our acceptanceof it is no longer the acceptanceof testimony as such;
it is the affirmationof something based upon evidence, that is, historical knowl-
edge. (Idea, 257)
Let us see what this really means. First of all, we must be careful not to
be bewildered by the statement just quoted through a possible confusion of
evidence and testimony. For many, the testimony found in our documents
and other written sources is just what we mean by the best possible sort of
historical evidence, though we recognize that there are other kinds of ev-
idence too. But in Collingwood's view the mature and autonomous science of
history itself determines what is or is not evidence. That some assertion is
transmitted over the course of time and exists preserved in writing in some
archive or library is of no interest at all to historians, in Collingwood's view,
until such time as it proves relevant to answering questions historians seek to
answer. The testimony preserved in the documents qua testimony, is not,
then, historical evidence. Its becoming historical evidence is entirely de-
pendent upon the historian and the use he has made of it. The point of the
quoted statement should now be clear. To be scientific - in the older,
broader sense of the term preferred by Collingwood, rather than in the
narrower sense more commonly used these days - historical propositions
must rest upon evidence, grounds for the admission of which are rooted in
the character of the inquiry itself. Assertions which rest upon testimony
depend upon the merely fortuitous or accidental, and that is hardly a firm
basis for scientific knowledge.
At first blush, we seem to face a paradox. It is no secret that history has
had its credentials as "objective" challenged much more than most disci-
plines, and if it is to have any hope at all of holding up its head in the
company of epistemically respectable disciplines that will surely depend on
its ability to base itself upon evidence which is in some sense given and
does not depend upon the vagaries - the subjective vagaries - of historians
and their questions. But if Collingwood is right, this support is no longer

there. The documents and testimonies which have been transmitted from
bygone days, far from being the touchstone in terms of which to determine
historical truth, now appear themselves to depend for their status on the
workings of an autonomous discipline of history. Yet a moment's reflection
makes it clear that the paradox is merely apparent, that it rests upon the
plausibility of some sort of common-sense realism and cannot survive its
rejection. It would not be appropriate here to open this question to discus-
sion; our concern, after all, is to make clear what Collingwood's views are
without detailed exploration of all the problems which would have to be
cleared up before one could claim to have a worked-out theory of historical
knowing. In any event, the anti-realistic line relevant to removal of the seem-
ing paradox is to urge that knowledge and knowledge claims do not rest
simply upon a direct interaction of knower and the known, but depend upon
acceptance of criteria in terms of which knowledge is delineated by those
jointly engaged in some knowing enterprise - the small number of special-
ists when that enterprise is the disciplined pursuit of scientific or theoretical
knowledge, the total community or the largest part of it when what is in-
volved is the every-day knowledge of common sense and social existence. It
is only in terms of such a conception that the historian's enterprise is possible
or intelligible, since the realistic interaction of knower and known is clearly
not possible in this sphere. It turns out, then, that what history in fact does is
precisely what Collingwood says a discipline must do if it is to be worthy of
the name of knowledge: it does not simply accept what has been fortuitously
preserved. Rather, its claims to knowledge are vindicated by appeal to
grounds. And that such-and-such are indeed the grounds for this-or-that
claim to historical knowledge is determined in terms of the criteria that the
discipline uses.
History is a responsible discipline, and the historian is not at liberty to
make up the facts as he needs them. Nor is a work of historical narrative so
self-complete and independent of everything else as to require no reference
to anything beyond the needs of its own inherent logic.38 If nothing could
serve as historical evidence, then clearly there would be no possible discipline
of history, and it is worth remembering that the main emphasis in the first
two-thirds of The Idea of History, the part concerned with the development
of the idea of history or the coming to maturity of its generic essence, is
the varied and changing ways in which historians deal with historical ev-
idence. But, in the end, it is the discipline which is the master; it is not
subservient to what has simply managed to survive. Nor is this particularly
38. For an opposite view, see W. B. Gallie, Philosophy and the Historical Under-
standing (London, 1964), chs. 1-5.
strange. Ti terms of what, after all, are we to decide that something is ev-
idence for something else, or that of two somewhat divergent accounts of
what purports to be the same occurrence one is better than the other? It is
clearly not in terms of what really happened when the past was present,
since that is not available for the historian's inspectionA9 And this leaves
only the discipline itself, the principles and criteria in terms of which it
assesses the cogency and acceptability of putative historical evidence. With
this in mind, we can now turn to the last of the three sections of The Idea
of History we are considering, "History as Re-enactment of Past Experi-
ence."
VIII
The section begins with the question, "How, or under what conditions, can
the historian know the past?" (Idea, 282). The direction of Collingwood's
thought makes it clear that the main condition is that there be something to
engage the historian now. The actual events which transpired when the past
was present are clearly no longer available. And we have already seen why
he cannot allow that it is sufficient for the historian to find testimonies or
other accounts in his sources, which accounts he may simply - acceptingly
copy in his own narrative. What, then, is left for the historian to do?
It is here that we get Collingwood's view that history is the re-enactment
of past thought. Thus: "The difference between memory and history is that
whereas in memory the past is a mere spectacle, in history it is re-enacted in
present thought" (Idea, 293).40 Taken out of context, this seems absurd, and
even in context it may well be incorrect. But it is at least intelligible when
one sees what Collingwood attempts to accomplish with it. It emerges as an
attempt to answer the question: Since history is being done and is, therefore,
patently possible, although the events of the past are forever beyond our
acquaintance, what is there for the historian to deal with? The answer turns
out to be thought. Without dealing in detail with Collingwood's view of
thought or thinking, it will suffice to point out that for him it is something
which can survive in the way that events - social events and psychological
ones - cannot. The historian understands what transpired when he is able
to re-think the thought of the historical actor he is dealing with. In some
relevant sense, Collingwood takes the actor's thought and the historian's re-
thinking it to be identical (Idea, 288 if.). In seeking to understand a theorem
of Euclid's Elements or the point of the argument in some dialogue of Plato,
39. See my "Evidence and Events in History," 175-179.

40. Cf. "If we raise the question, Of what can there be historical knowledge? The
answer is, Of that which can be re-enacted in the historian's mind" (Idea, 302).
to use examples he uses himself, one has, in effect, to re-think their thoughts.
There is no significant sense in which the original thought and its present
re-thinking are two, not one.
Before bringing this long discussion of Collingwood's theory of historical
knowing to an end by completing our account of what is involved in this
idea of re-thinking, we had better be sure we fully understand how it works.
It may seem to have some point in intellectual history - understanding
Euclid or Plato -but there is more to history than that. Collingwood, him-
self, says that politics, warfare, economic activity, and religion are among the
proper subjects of historical inquiry (Idea, 309 ff.), but I dare say that any
number of his readers have failed to be convinced that, given his views, he
could make that out. I do not wish to make up arguments which might be
offered in a Collingwoodian spirit in support of these assertions. Rather, I
prefer to turn to an actual instance of Collingwoodian history, written by
Collingwood himself. Though Collingwood as historian worked on the
seemingly unlikely brute realia of archaeology, his way of dealing with them
is quite consistent with his conception of historical knowing as present re-
thinking and gives substance to his claim (Idea, 296) that one recovers the
thought of others with the aid of evidence.
As is well known, Collingwood combined a career as philosopher with
that of historian, being for some time University Lecturer in Philosophy and
Roman History at Oxford (Roman, Vii).41 His own special field was Roman
Britain, and for a time he did quite a bit of writing on the subject. I do not
know what his reputation was among his fellow historians,42 but one may
suspect that his work was held in some esteem, since he was invited to
contribute the more than three hundred pages devoted to Roman Britain
in the Oxford History of England. But whether esteemed or not, his historical
work does show that he himself felt no difficulty about the possibility of
writing history which was anything but intellectual history or the history of
ideas. And if the sort of history he writes can be treated as being the re-
thinking of past thought, it seems only reasonable to conclude that that is
what he meant by his famous phrase all the time. It seems considerably
more sensible to explicate the phrase in the manner here suggested than to
determine what it means by a contextless analysis of it as simply a phrase in
the English language.
One of the major preoccupations of Collingwood's Roman Britain is to
41. Page references given in the text preceded by Roman will be to Collingwood's
"Roman Britain" in R. G. Collingwood and J. N. L. Myres, Roman Britain and the
English Settlements, 2nd ed. (Oxford, 1937).
42. Cf. I. A. Richmond's generally very favorable "Appreciation of R. G. Colling-
wood as an Archaeologist," Proceedings of the British Academy 29 (1943), 476-480.
I am grateful to W. von Leyden for calling this to my attention.
make sense of archaeological remains; he seems particularly interested in
fortresses, walls, and the like. The question frequently arises: "why should
anyone build something like that?" and, of course, the attempt to provide an
answer leads Collingwood to try to discern a purpose that the construction
could serve. Although I do not wish to repeat Collingwood's detailed descrip-
tions of the remains he deals with-which are hard to visualize in any
event, since the volume has no plates or illustrations-I do think that one
good example will make clear what is involved.
Somewhat to the south of Hadrian's Wall, but following the line of the
Wall, there is an earthwork which, Collingwood tells us, English antiquaries
call the Vallum (Roman, 124). After presenting an account of its chrono-
logical relations to the building of the Wall itself, the question of why it was
built is dealt with, and we are told that this has always perplexed scholars
who have formulated hypothesis after hypothesis only to have them rejected
as incompatible with the actual features of the Vallum (Roman, 133).
Collingwood proceeds to detail some of the features in order to determine
what it might or might not be used for. We are told (Roman, 133),
that the Vallum in its original shape was a formidable obstacle to traffic, but
incapable of military defense, and so designed, indeed, as to look ostentatiously
unmilitary;that this obstacle is carried with remarkablethoroughness, admitting
no interruptionwhether from hard rock subsoil, morass, or ravine, right across
the country from Tyne to Solway, close behind the Wall; that, according to the
latest results of excavation, it was made at the same time as the Wall itself; and
that the only original ways across it are solid causeways opposite the Wall forts
and perhaps also opposite the milecastles, each surmounted by a stone gateway.
In sum, the Vallum is a second obstacle parallel to the Wall and provided with
a correspondingseries of controlled openings for traffic, differing from it in its
deliberatelyunmilitarydesign.
If at first blush a second obstacle seems a redundancy and one is, hence,
tempted to think of it as an earlier, perhaps a more primitive, less satisfactory
attempt to accomplish what was better done subsequently, we are forced to
face both the archaeological evidence of their contemporaneity and that the
unmilitary character of the Vallum would presumably preclude its serving
the defensive function the Wall of Hadrian is known to have had.
The question, then, is "what function could the Vallum serve?" - or,
indeed, "what is the Vallum?" Drawing upon what he knows of Roman
frontiers, administrative arrangements, and the like, Collingwood offers an
account concerning which he says, "There is no proof that this explanation
of the Vallum is correct. All that can be claimed for it is that it fits the
facts" (Roman, 134). Yet it is so excellent an example of the kind of his-
torical thinking one finds throughout Roman Britain that it is worth quoting
in full the three paragraphs involved (Roman, 133 f.).
Now, a Roman frontier has two functions, one military or defensive, the other
financial, as a line where traffic passing between the province and the uncon-
quered country outside it passed through supervisedopenings and paid duty. And
it is a peculiar feature of Roman administrationthat the financial service un-
der the procuratorswas entirely separated from the military service under the
provincial governors. The sentry on guard at the gate of a fort was responsible
to the commandant;he, whether or not through the mediation of a legionary
commander,to the governor of the province; and he to the emperor. A customs
officer was responsibleto the procurator,and he to the emperor directly, without
any interventionof the governor. And not only were these two services separate,
but relationsbetween them were delicate: friction and jealousy were not unknown.
Before the building of Hadrian's Wall, continuous frontier works, where they
existed at all, had been structurallyseparate from the forts of their garrisons.
It may be conjectured (we have no proof) that the openings in the barrierwere
controlledby customs officers,while obviously the sentries of the garrisonslooked
after their own fort gateways. But on Hadrian'sWall the forts, with one or two
exceptions, formed part of the barrier itself, so that a man passing through the
fort was passing the line of barrier; and (again with one or two excep-
tions) there was no way of crossing the barrier except by going through a fort,
unless, indeed, non-militarytraffic was allowed to pass through milecastles; if it
was, the same problemwould arise there.
From a militarypoint of view, this new method of planning the forts in relation
to the barrier was no doubt an improvement.If traffic crossing the line of the
barrierwas compelled to pass through a fort, the military control over such traf-
fic was tightened. But the question must now have arisen, how to provide for the
customs officers?Hadrian, a stickler for military discipline, may very well have
thought it unwise to give the procurator'sman an official position at fort gate-
ways, where the authorityof the commandantshould be undisputed.The simplest
solution on paper, though a cumbrous and expensive one, would be to have a
second barrier behind the Wall; to make this barrier look as unmilitary as pos-
sible, consistent with efficiency; and to provide it with a crossing opposite each
fort, where the customs officers could do their work. The Wall as a whole would
be controlled by the governor, the Vallum by the procurator;the distinction be-
tween the two reflectingand symbolizingthe separationbetween the military and
financialservices.
As I have indicated, this kind of thinking is typical of what Collingwood
does in this study. I think by itself it will prove sufficient to help us elucidate
what he means by the historian's re-thinking.43
Since The Idea of History is not part of any actual instance of historical
research, at least not in the sense of an historian's reconstructing some past
event,44 what Collingwood can mean by some of his assertions may seem
43. For other instances of Collingwood's historical thinking see Romanl, 140, 142,
159, 160 (Postscript), and 172 f., as well as his extended discussions of the strategy
of Julius Caesar, Claudius, and Aulus Plautius.
44. Of course, the first two-thirds of the book is an historical sketch of the develop-
ment of the idea of history, but that is not the sort of work I have in mind. In any
event, the sections of the last part of the volume were separately written essays or
lectures.
perplexing at times. He has been telling us that history is the re-enactment of

past thought, and then says:
The historian cannot apprehendthe individual act of thought in its individuality,
just as it actually happened.What he apprehendsof that individual is only some-
thing that he might have sharedwith other acts of thought and actually has shared
with his own. . . . It is the act of thought itself, in its survival and revival at
differenttimes and in differentpersons: once in the historian'sown life, once in
the life of the person whose history he is narrating. (Idea, 303)
One can imagine philosophical critics wondering about what sort of thing an
act of thought is which is to be apprehended by another who can share it
"itself" but cannot apprehend it in its "individuality." But it does not really
matter. Whatever these sentences may be thought to mean, or whatever
someone else might mean who through some extraordinary coincidence hap-
pened to express himself in these very same words, what Collingwood means
to convey by them is clear enough in light of his account of the Vallum and
its purpose. Collingwood does not claim - indeed, he denies - that he is
able to experience - in whatever sense is suitable here - the actual act of
thinking that took place in antiquity.45His claim is that he can reconstruct
the thought behind the decision to build the Vallum, which means, of course,
that he is able to re-think the thought - though not, to be sure, exactly in
the same order in which the various elements of the project were taken into
account, and not in Latin but in English. The problem of the Vallum cannot
be solved until the historian concerned with it can figure out what human
purpose it served or could serve. What that purpose is, is not self-evident;
one does not immediately hit upon it as soon as one determines its physical
character. One has to think about what could be done with it, and not by
one of us but, rather, by a Roman official highly enough placed as to be
able to have such a thing made.46 If Collingwood's solution to the Vallum
45. Cf., "To return to our supposed objector. Why did he think that the act of
thought, by becoming subjective, ceased to be objective? . . . It is because he under-
stood by subjectivity not the act of thinking, but simply consciousness as a flow of
immediate states . . . . What the objector was doing, therefore, was to assume that
all experience is immediate, mere consciousness, devoid of thought" (Idea, 294). It is
thought, not immediacy, which Collingwood takes to be the object of historical knowl-
edge.
46. One has to put oneself in Hadrian'splace and make sense of what he did in terms
of the problems he was seeking to deal with, the questions he sought to answer. Cf.
C. V. Wedgwood's criticism (in The Sense of the Past [originally published as Trutih
and Opinion], Collier Books ed. [New York, 1967]) of Gibbon's condemnation of the
actions of Sulpicianus after the murder of his son-in-law, the Emperor Pertinax, as
reflecting Gibbon's inability to see the problems from Sulpicianus' own standpoint:
"Gibbon cannot and does not imagine himself in the predicament of Sulpicianus and
an obvious element in the situation therefore escapes him. Gibbon's method enables
him to describe and explore the surface of events with incomparable brilliance, but it
rarely leads to any penetration below the surface" (34). Clearly, Miss Wedgwood's
problem is correct, there is a clear sense in which he has re-thought the

thoughts of Hadrian in all of their historically-relevant character. The essen-
tial considerations which presumably passed through Hadrian's mind as he
came to the decision to have the Vallum built have passed through Colling-
wood's as well. There is no suggestion of his having entered fully into the
existential experience of the historical actor in the sense of reproducing the
feelings, emotions, and other appurtenances of existing-and-experiencing-
here-and-now in the way that an historical novelist might seek to do.47 Here,
it seems, is the central feature of what Collingwood thinks history is: it is re-
thinking thought on the basis of evidence without ever becoming psychology.
As re-thinking thought, its object is what can be detached from the original
context of action and be reproduced in the later context of historical inquiry,
hence its object is universal and not existential.48
We are now in a position to offer an account of what is involved in one
of the most quoted of Collingwood's assertions: "For history, the object to
point is that Gibbon simply failed to re-think Sulpicianus' thought in Collingwood's
sense. Note also her following remark in an entirely different context: "The wars of
religion have left their mark on the institutions, the society and the prejudices of most
of the peoples of western Europe. These bitter and distracting conflicts are not only
dismal and deplorable but largely incomprehensible until we can bring to them some
understandingof the beliefs which guided the protagonists" (46).
47. Hugh J. Schonfield's fascinating and suggestive The Passover Plot (New York,
1966) is frequently flawed by just such a pretense that he is able to write or speculate
about how Jesus felt at some time or what he might have said at another.
48. It should be clear from all this that when Collingwood says that historians re-
think past thought he means to have "thinking"taken literally. But the sheer incapacity
of his critics to recognize thinking taken seriously as thinking taken seriously is, indeed,
a wonder to behold. Thus, Gardiner says that "the suggestion of some sort of telepathic
communication with past thoughts is too insistent to be entirely disregarded" (op. cit.,
39) and that "the subject-matterof history is made to appear very mysterious indeed,
demanding tentative handling and esoteric methods" (ibid., 49). F. D. Newman speaks
of "Collingwood and his infamous doctrine of empathic understanding" (Explanation
by Description [The Hague and Paris, 1968], 51; the text actually says "emphatic,"
but this is clearly a typographical error). And W. H. Walsh proclaims that "It is not
true that we grasp and understandthe thought of past persons in a single act of intuitive
insight" (An Introduiction to Philosophy of History [New York, 1951], 58) as if Colling-
wood ever held that it was. (Walsh has discussed Collingwood's views elsewhere as
well: "R. G. Collingwood's Philosophy of History," Philosophy 22 (1947), 153-160, and
"The Character of a Historical Explanation," Aristotelian Society Supplementary Vol-
nlne XXI (1947), 51-68. One occupational hazard of doing philosophy is that your
words may be given the worst possible interpretation by a critic who then saves the
day by claiming what you had been trying to say all along. The treatment of Colling-
wood in Walsh's book exemplifies this very nicely.) Other such critics include L. Jona-
than Cohen, "A Summary of Work in the Philosophy of History, 1946-1950," Philo-
sophical Quarterly 2 (1952), 173-177, and "Has Collingwood been Misinterpreted?'
ibid. 7 (1957), 149-150 and G. J. Renier, History: Its Purpose and Method (Boston,
1950). How these direct, immediate, non-thinking forms of mental communication have
been foisted upon a writer so unambiguously committed to thinking must be one of
the most remarkable intellectual accomplishments of our time.
be discovered is not the mere event, but the thought expressed in it. To dis-
cover that thought is already to understand it. After the historian has ascer-
tained the facts, there is no further process of inquiring into their causes.
When he knows what happened, he already knows why it happened" (Idea,
214). Actually, it is only the last of these sentences which is really famous,
but I have chosen to quote a somewhat larger extract because my interest
here is different from that of any number of others who have discussed these
words. For the most part, they are discussed in writings on historical ex-
planation, but my interest here is not in Collingwood's views on that subject,
but, rather, on his conception of historical knowing. Hence, I begin the
extract quoted with Collingwood's statement of what the object to be discov-
ered is -that is, the object to be discovered no matter how it is to be
explained. I should not wish to deny that the last of the sentences quoted
contains a conception of historical explanation, but it is a very unusual
sentence. It takes all of its words to express its conception of historical
explanation, yet it manages to say something else as well, something about
the nature of "what happened," that is, about "the object to be discovered"
in historical research.
It should be possible to provide a complete description of that earthwork
which is the Vallum in physical and chemical language. Likewise, in prin-
ciple, as the saying goes, an ideal observer49 or a whole research team of
ideal observers might have been able to provide a complete behavioristic
description of the movements of Caesar's forces in the course of his invasion
of Britain. But the former would not be an account of the Vallum qua the
Vallum, and the latter would not be a description of an invasion qua inva-
sion. The Vallum as such is a product of human purpose; without that
purpose it is something simply there, and unintelligible. Consider the follow-
ing well-known example: "Suppose that I enter a bank, I then take a with-
drawal slip and fill it out, I walk to a teller's window, I hand in my slip, he
gives me money, I leave the bank and go on my way. Now suppose that you
have been observing my actions and that you are accompanied by, let us say,
a Trobriand Islander. If you wished to explain my behavior how would you
proceed?"50Presumably, the Trobriander sees the course of what takes place
as a string of unintelligible bits of behavior. The other observer sees it as
an instance of withdrawing money from one's savings account. The problem
is to get the Trobriander to see it in the same way, so that when, at last, he
knows what happened, that money was being withdrawn, that each of these
discrete - describably discrete - bits of behavior are parts of a certain
49. Who, of course, would not be an historian.

50. Maurice Mandelbaum, "Societal Facts," in Theories of History, ed. Patrick Gar-
diner (Glencoe, Ill., 1959), 479.
course of action, he also knows why each describably discrete bit of behavior
happened. When Collingwood knows what the Vallum really is, he knows
why it was built. When he knows what Caesar's strategy was, he knows the
record of the ideal observer noted above not as behavior, but as the move-
ment of troops. In sum, what is involved in the passage quoted from Colling-
wood is simply the claim that the object of the historian's research is human
action, not behavior understood physicalistically.51
This is no place to raise the issues of behaviorism. Whether behaviorism
is tenable or not, Collingwood's point is that its object is not that of the
historian. If all that we could have were the observations of behavioristic
psychologists - if, that is, the methodological requirements of behavioristic
psychology became the elements of the only possible theory of human
nature52 then, in Collingwood's view, history as a discipline would be im-
possible. The object of historical research is human thought, because only
thought can be abstracted from its original context and reproduced in the
historian's mind. What Collingwood is trying to say is that the historian
studies action - an "outside" informed by an "inside" - not behavior. Any-
thing exhausted by its immediacy cannot be reproduced in the mind of the
historian.53
51. Critics (cf. Walsh, An Introduction to Phlilosophly of History, 52) have been
quick to cite Collingwood's "When he knows what happened, he already knows why it
happened" (Idea, 214) as evidence of his intuitionism, and taken out of context, it does
seem to express intuitionist sentiments. But we have already seen that Collingwood was
not and could not be an intuitionist, a conclusion surely reinforced by my non-intui-
tionist interpretation of the offending passage. (Actually, the argument really goes the
other way. It is precisely because Collingwood cannot be an intuitionist that seemingly
intuitionist assertions require to be given non-intuitionist interpretations. We are not,
after all, dealing with contextIess statements in ordinary English.)
52. See my "The Phenomenological and Naturalistic Approaches to the Social," in
my ~~~~~b
Phlilosophly of the Social Scienices: A Reader, ed. M. Natanson (New York, 1963),
289 ff.
53. Collingwood's use of the "inner-outer"metaphor to characterize human action
has provided another occasion for critical attack. All he intended by the expression is
that human action consists of behavior informed by thought, and it is difficult for me
to see how any fair-minded reader could fail to see that. Gardiner says that the expres-
sion is "artificialand misleading": artificial because we do not talk that way about hu-
man action - a petty observation of no importance-and misleading because "the in-
troduction of a spatial metaphor gives the impression that what are called 'insides' of
events are queer objects, invisible engines that make the wheels go round" (ibid., 47).
To say that Collingwood has used a metaphor that may be misleading is both to recog-
nize it for what it is and not to be misled by it. But what follows in Gardiner's book
is an attempt to show that if anyone was misled it was Collingwood himself; or per-
haps "inside-outside"is not a metaphor after all. I would only say that we have already
found in Gardiner (see notes 20, 21, and 28) an inclination to convert poor Colling-
wood into the one thing he would not be, an historical realist believing in the existence
of strange objects which the historian knows directly by intuition or acquaintance. This
is simply another instance of the same tendency. Given what Collingwood's actual views
were, he could never have intended that the metaphor be taken literally and that human
I.x
The burden of Collingwood's investigation into the nature of historical

knowing is to discover the grounds of the possibility of knowing now what
thought be something spatial. (Indeed, I cannot imagine anyone of any philosophical

persuasion not bereft of sense believing in the various entities - pasts which are present,
thoughts which are spatial, and the like - that Gardiner attributed to Collingwood).
Gerd Buchdahl's treatment of the "inner-outer" matter is also worth our attention
("Logic and History, An Assessment of R. G. Collingwood's Idea of History," Auastral-
asian Journal of Philosophy 26 [1948], 94-113). Collingwood notes that for "the sci-
entist, nature is always and merely a 'phenomenon,'" (Idea, 214) but though he uses
"phenomenon" there is no reason to think that he has Kant in mind. For he immedi-
ately explains what he means: "not in the sense of being defective in reality, but in
the sense of being a spectacle presented to his intelligent observation; whereas the
events of history are never mere phenomena, never mere spectacles for contemplation,
but things which the historian looks, not at, but through, to discover the thought within
them." The point is two-fold: human beings act, not simply behave, and one must get
to the thought to know what the act is. And given the pastness of the past, its non-
observable character, it is this, the possibility of getting to the thought, that makes his-
torical knowledge possible-assuming, that is, that thought is what Collingwood takes
it to be and his belief that the same thoughts may be thought by many. But in Buch-
dahl's treatment all this is lost or ignored. Instead the distinction between inner and
outer becomes Kant's distinction between things-in-themselves and appearances. Thus,
Collingwood "allowed himself to be misled by the metaphysical use of the adjective
'inner,'" and owing to his preference for internal relations - a doctrine Buchdahl foists
on Collingwood from Bradley on the ground that "Like most philosophers, Collingws~ood
is not too clear on the nature and content of his ideals, is not clear how what he feels
as a deficiency determines the choice of his ideals" ibidd., 96) -he thought that thought
presumably the historical thing-in-itself, would enable him to have them. Having sad-
dled these incredible views upon Collingwood, Buchdahl now criticizes them: "But it
seems to me that in the end the other person's thought, particularly when this thought
is reconstructed, is just as much reached from the outside . . . is as much an
appearance' as anything else. That Collingwood did not realize this may be due, among
other things, to his peculiar views of thought . . ." ibidd., 98).
Buchdahl's treatment of Collingwood is altogether an incredible performance. To
understand Collingwood, he finds hints in Bradley and others in Kant; what he does
not do is allow Collingwood's own concerns, together with the way the actual discipline
of history is practiced, to provide the context in terms of which to interpret The idea
of History. This will be clear from a moment's consideration of the way he brings in
the doctrine of internal relations mentioned parenthetically in the previous paragraph.
Making the usual realistic assumptions, Buchdahl thinks that Collingwood "loosened the
facts from a solid background" and was then confronted by the need to tie them back
together in an even more intimate way than they were before. Collingwood's motive is
said to be that he is committed to the doctrine of internal relations, and it is that which
led him to emphasize that thought is the material of history: "'Thought' you will find
is the model making the assertion of internal relations plausible" (ibid., 96). But all
this is simply wrong. It is only because he has not paid any attention to how the his-
torical past is constructed by historians that Buchdahl can say that Collingwood loos-
ened the facts from solid moorings: as if the historian ever confronts real, solidly-
moored historical facts, and as if he does not himself construct his facts on the basis
of historical evidence. As for thought, we have already seen how, far from being intro-
duced to justify the Bradleyan doctrine with which Buchdahl saddles him, its central
place in Collingwood's theory is entirely owing to his view that it alone can survive its
original context to be re-enacted in the historian's present.
happened then. That historical knowledge is possible, Collingwood did not

doubt. His Speculum Mentis view was that history is concerned with a world
of fact. The holistic view of knowledge he then held led him to conclude
that historical knowledge was not possible, but even without that, the ob-
jectively-there character of that world of fact would have led him to reject it
as soon as his anti-realistic views crystallized. He could not, however, simply
deny history, and seeking to discover just what it could possibly be about
what subject matter one could find for it to deal with - he came to think of
it as the science of historical evidence. Any number of essays between
Speculum Mentis and The Idea of History take this view. But while much is
made of the critical treatment of historical evidence, that middle position is
not fully satisfactory inasmuch as it fails to account for the historian's treat-
ment of events. Without some way of bridging historical evidence and his-
torical events one of the central concerns of the historical enterprise remains
unintelligible. His quest for the solution of this problem, beyond doubt ac-
companied by very serious attention to what he was himself doing as an
historian, led him back once again to a world of historical fact, but not to the
lived experience of the historical agents he was concerned about. Packed into
The Idea of History is a conception of action and a conception of thought,
which together hold that human action is behavior informed by thought, that
history is concerned only with action, and that by making proper use of
historical evidence the historian can bring to his own mind thoughts once
thought before. History can be about only what the historian can know: not
the real past, but thought he can re-think.
I do not wish to say that Collingwood's most mature doctrine is free of
problems, though it is always fair to remember that illness and premature
death prevented his bringing his philosophy of history to a conclusion. In
particular, even if he is right in thinking that history can only be concerned
with what is informed by thought, to be informed by thought is not to be
exhausted thereby. One still requires a non-realistic account of the relation
of the entire historical event that the historian reconstructs to historical ev-
idence.54 But even as far as we have it, the doctrine of The Idea of History
is certainly not the work of an historical skeptic, and only one who is either
a realist in epistemology or a behaviorist in philosophical anthropology could
credit that charge against it.
State University of New York

at Binghamton
54. Cf. my "Evidence and Events in History."

Goldstein Collingwood's Theory of Historical Knowing

Uploaded by

Copyright:

Available Formats

Goldstein Collingwood's Theory of Historical Knowing

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Goldstein Collingwood's Theory of Historical Knowing

Uploaded by

Copyright:

Available Formats

Wesleyan University

Collingwood's Theory of Historical Knowing

In his review of The Idea of History, Maurice Mandelbaum asserts that

1. Maurice Mandelbaum, review of R. G. Collingwood, The Idea of History, in

7. Errol E. Harris, "Collingwood on Eternal Problems," Philosoplvical Quarterly 1

12. A term Collingwood uses in An Essay on Philosophical Method (Oxford, 1933).

of The Idea of History is to present an account of the gradual overcoming

I amntrying in the present discussion to provide a context in terms of which

16. Indeed, it already contains a version of Collingwood's anti-realistic view of know l-

lows is an explicit denial that what he said is skeptical; "for scepticism

statuses. By "the facts of history" Collingwood would understand facts estab-

that it is autonomous, that it is not dependent upon things outside of itself,

pendence. I have just spoken of the autonomy of history, but he writes of

The relation of history -- the discipline - to historical evidence is one of

and their questions. But if Collingwood is right, this support is no longer

39. See my "Evidence and Events in History," 175-179.

perplexing at times. He has been telling us that history is the re-enactment of

problem is correct, there is a clear sense in which he has re-thought the

49. Who, of course, would not be an historian.

The burden of Collingwood's investigation into the nature of historical

thought be something spatial. (Indeed, I cannot imagine anyone of any philosophical

happened then. That historical knowledge is possible, Collingwood did not

State University of New York

54. Cf. my "Evidence and Events in History."

You might also like