[forthcoming in Australasian Journal of Philosophy]

Allais, Lucy, Manifest Reality: Kant’s Idealism and his Realism, Oxford: Oxford University Press,
2015, pp. viii + 329, £40 (hardback).

In this ground-breaking book, Lucy Allais offers a proposal for interpreting Kant’s
transcendental idealism that draws on contemporary ‘direct realist’ accounts of perception. The
result is a compelling piece of Kant scholarship that also offers a wider readership new ways of
thinking about the Kantian philosophical legacy.
The meaning and import of Kant’s transcendental idealism has been debated since the
publication of the first edition of the Critique of Pure Reason in 1781. It informs nearly everything
that Kant does in the Critique, both positive and negative; and it bears on his handling of
freedom in his ethical works. But most fundamentally transcendental idealism is a thesis about
space and time. Transcendental idealism denies the absolute, mind-independent, reality of space
and time. Space and time are not self-subsisting entities; they are not any sort of determination,
property, or relation proper to things as they are in themselves, i.e. apart from these things’ being
intuited or sensibly represented [Critique of Pure Reason, A23/B37, A39-40/B56-7]. Space and time
are ‘pure forms’ of human intuition: anything that can be sensibly represented by us must be
represented in space and time.
Transcendental idealism has contested epistemological and metaphysical implications,
which Allais outlines as follows [4]:

(1) Kant distinguishes between things in themselves … and things as they appear to
us…. This is closely related to, but not exactly the same as, his distinction between
noumena and phenomena.
(2) Kant argues that the spatio-temporal objects of our experience … are mere
appearances, or mere representations that do not exist apart from a connection to
possible perceptions.
(3) Kant claims that we do not and cannot have cognition … of things as they are in
themselves.

Each claim has been interpreted in a range of ways over the years, yielding wildly divergent
interpretations of transcendental idealism. At one extreme are the ‘deflationary’ interpretations
brought to prominence with Henry Allison’s 1983 Kant’s Transcendental Idealism. For deflationists,
the distinction between appearances and things in themselves is epistemological, or perhaps
methodological — not ontological. There is no commitment to two kinds of entity, but rather
only to one kind of entity — the appearances that we can cognize — though we can have a
merely notional thought of these things as they are in themselves. On the other end are ‘extreme
metaphysical’ readings that have come to greater prominence in the past couple of decades: a key
contribution is Rae Langton’s 1998 Kantian Humility: Our Ignorance of Things in Themselves. Such
interpreters take Kant to be committed to the existence of noumena, conceived in some positive
terms (e.g. as having causally inert intrinsic natures). Another set of interpretive questions arises
about the idea that we know objects only as appearances: chiefly, whether this commits Kant to
a phenomenalism whereby objects of experience are nothing other than constructions from mere
mental representations.
Allais aims to establish a ‘moderate metaphysical’ position between the ‘deflationary’ and
‘noumenalist’ extremes, and roundly rejects phenomenalism. On the one hand, she wants an
interpretation that does full justice to the many places and many ways that Kant acknowledges
that there is a way things are in themselves, independently of the conditions of human cognition:
this is her nod to the metaphysical side. On the other hand, she wants an interpretation that does
justice to Kant’s many remarks about the mind-dependence of appearances, without falling into
phenomenalism. Here Allais elaborates the empirical realism that Kant took to be the flip side of
transcendental idealism, a move that amounts to something of a nod to the deflationary side.
Allais’s crucial interpretive lesson consists of showing that the conclusions one draws
about transcendental idealism will vary considerably depending on what one assumes as a theory
of perception. This is a brilliant move that throws the long slog through the secondary literature
(which occupies the first four chapters of the book) into sharp new relief. Kant, she points out,
did not have a ‘theory of perception’ in any contemporary sense, and it is not her project, exactly,
to reconstruct one. She argues that Kant is committed to rejecting what she calls a ‘Cartesian’ or
‘indirect’ model of perception, whereby we perceive objects through mental intermediaries and
‘the object itself is not a constituent of the mental state a subject is in when perceiving’ [12].
Having argued in Chapter 2 that the preponderance of textual evidence tells against Kant’s
commitment to phenomenalism, Allais then proposes that direct realist accounts of perception
allow us to accommodate Kant’s views about the mind-dependence of appearances without
reading transcendental idealism in phenomenalist terms [103].
In the Critique’s Refutation of Idealism, Kant argues that our mental representations put
us in direct contact with external objects, and indeed that we stand to know our own minds only
in such relations. Allais begins her positive story by acknowledging the friendliness of these
claims to direct realist views of perception, and more specifically to relational views [102; see also
135]. According to direct realism, ‘what is directly present to us in perceptual consciousness is
the world of objects outside us’ [110] — rather than mental intermediaries that we then must
interpret in order to represent objects. Relational theorists add that the objects themselves are
constituents of perceptions. Allais draws on this to understand Kant’s claims about empirical
objects being mere appearances. What figure in appearances are what Allais refers to as
‘essentially manifest qualities’. A model for this comes from certain accounts of colour, where
(e.g.) redness is essentially and irremediably a way a thing might look. For Allais’s Kant, this
generalises: all that can be given in appearances are the essentially manifest qualities of things
[125]. Allais draws here on a passage from the Prolegomena where Kant says that ‘all the properties
that make up the intuition of a body belong merely to its appearance’ [Prolegomena 4:289]. This passage
has met with vexed reception since, as Allais acknowledges, Kant elsewhere rejects the idea of
explaining idealism through any kind of comparison with colour [Critique B45]; and plausibly the
Critique should be given the upper hand in any textual duel with the Prolegomena, which Kant
wrote only to popularize the Critique. But Allais deftly argues that what is at stake at Critique B45
is a conception of secondary qualities as ‘mere alterations of our subject’; Kant only rejects that
such conception could guide an interpretation of transcendental idealism. This leaves open the
possibility that Kant could embrace a model of secondary qualities as essentially manifest, in
terms of a relational theory of perception [127]. Allais concludes that making sense of the mind-
dependence of appearances together with the Refutation of Idealism requires a relational theory
of perception [135].
Readers of the Critique might think that direct realism is a non-starter for Kant, given all
of his talk of ‘synthesis’ — talk that is, perhaps, more readily interpreted through an indirect or
‘Cartesian’ account of perception. Allais has several avenues of reply here. Initially, she draws on
some of John Campbell’s metaphors for the relational model: that we are (as it were) looking
through a transparent medium to the objects themselves; and that this transparent medium
might be volatile and in need of constant adjustment — so that the cognitive processing might
lie in those adjustments that enable us to see clearly how things are. Cognitive processing, in
other words, need not be a construction out of actual or possible mental states.
Another avenue of reply leads to some of the most controversial claims in the book. In
short, Allais suggests that there are all sorts of cognitive processing involved in perceptual
experience, not all of which is what Kant calls ‘synthesis’ [109]. The ordering of sensible stimuli
in space and time binds it in certain ways that Allais takes to be fundamentally non-conceptual:
this is not synthesis, which Kant indicates requires the categories. Now, for generations, it has
been thought that the upshot of transcendental idealism is precisely a rejection of the sort of
‘givenness’ about the phenomenal objects that would result if Allais is right about non-
conceptualism: that is why it is controversial. On her side, Allais arguably has the many places
where Kant says that intuition puts us in direct, unmediated, relation to objects, and that objects
are given in intuition. She also has a (fiercely contested) passage where Kant says that ‘objects can
indeed appear to us without necessarily having to be related to the functions of understanding’
[Critique A89/B122]. But there are serious problems with this view, as it implies that we may
represent objects in intuition but yet without representing these objects as objects. It is not
entirely clear what this might mean. It also not clear how Allais could then succeed in holding on
to direct realism: for if intuitions must be interpreted through categorial syntheses in order to
have referential thought about objects (see Chapter 11), how then are the merely aesthetic unities
given in intuition not mental intermediaries of some kind?
Do we need non-conceptualism? Allais contends that we need it in order to understand
how Kant’s argument for transcendental idealism works — one that rules out the infamous
‘neglected alternative’ that space and time might be forms of our intuition and also aspects of
mind-independent reality [195]. The premise of Kant’s argument, as Allais reconstructs it,
requires the non-conceptualist premise that ‘an intuition is a representation which presents its
object to consciousness’. Thus what pure intuition represents is also present to consciousness.
But an independent object can only be present to consciousness by affecting us; and pure intuition
does not involve anything affecting us. Therefore pure intuition cannot present to consciousness
anything that ‘exists independently of our representing it’; so space and time are not themselves
any mind-independent feature of reality, nor do they present us with any such features [195].
I have focused mostly on one strand of Allais’s enormously rich and thought-provoking
work: the other side of the story concerns Kant’s commitment to there being a way things are in
themselves. I found this to be the less compelling part of her account. She repeatedly suggests
that, according to Kant, a coherent account of appearances requires an ontological commitment
to things as they are in themselves as grounding appearances. The basis for this was not
explicated head on until near the end, in an account of the Critique’s Amphiboly (Kant’s polemic
against Leibniz). Allais claims that Kant takes the idea of ‘there being relations without there
being something non-relational’ to be non-sensical [240]. But the Amphiboly is concerned with a
set of ‘concepts of reflection’ that does not explicitly include the relational and the non-
relational: thus there is more interpretive work to be done if Allais wants to rest her ‘moderate
metaphysical’ conclusions on this text.
All told, this impressive book commands us to rethink entrenched orthodoxies in the
interpretation of Kant. It ought to shape the discussion for years to come.

Melissa McBay Merritt