Consciousness

Chapter

Six

The unity

Unlike illusionist accounts of conscious-
ness, IIT is based on the idea that we can
have complete confidence about our own
conscious experiences: ‘the existence
of one’s consciousness and its other
essential properties is certain’, Tononi
(2015) claims, and from there the theory
proceeds to make inferences about the
causes of consciousness from that sup-
posedly secure basis.

‘Consciousness is unified: it is irreducible
to non-interdependent, disjoint subsets
of phenomenal distinctions’, says Tononi
(2015). At the same time, it is extraordi-
narily informative, appearing to contain
countless infinitesimally small chunks of
information. Yet what makes a conscious state informative is not how much infor-
mation it may or may not contain, but the fact that it is just one of potentially
billions of other possible states. Consider a human and a photodiode facing a
blank screen that is alternatively on or off. The photodiode can make only two
distinctions: ‘light’ or ‘dark’. The human can distinguish the light screen not just
from the dark screen, but also from a red and a green screen, and from other
screens, showing any number of films, and from sounds, thoughts, and so on. This
is a vast amount of information, for you can discriminate between all these states
and each state has different behavioural consequences. So, you are vastly more
conscious than the photodiode.

The theory proposes that each of the five essential properties of experience must
be accounted for by a corresponding causal property of the physical system – the
brain. The axiom of integration, or unity, means that every part of the system
must be able to both affect and be affected by every other part, because other-
wise the integration would be reducible to a subsection of the system. According
to IIT, ‘experience is a maximum of intrinsically irreducible cause–effect power’
(Tononi, 2015). This irreducibility is measured as integrated information, referred
to in the theory as phi, Φ, and this is calculated according to a series of mathe-
matical formulae which you can find in Tononi’s 2015 description. This means
that consciousness can be graded rather than being all-or-none. The specific ana-
tomical location of the neural substrate of consciousness is not yet specified by
the theory, but Tononi states that whether it turns out to be distributed among
most cortical areas or only a subset of them, and whether it includes all cortical
layers or only particular cell types, ‘IIT predicts that in each case the neural sub-
strate of consciousness should be a local maximum of information integration’
(Tononi, 2015).

Tononi suggests that IIT allows the possibility of zombies because there could
be systems which look identical to humans from an external perspective, but
whose physical substrate consists of lots of mini-complexes of a low maximum
Φ value rather than forming a large complex of high maximum Φ. Physical tran-
sistors in a computer are not like neurons because they cannot be grouped into
macro- elements with irreducible structures. ‘Hence the brain is conscious and the

‘consciousness varies

with integrated

information’

(Tononi and Koch, 2015, p. 15)

Intrinsic existence Composition Information

Integration Exclusion

blue book

book

left blue

FIGURE 6.4 • Axioms of Integrated Information

Theory (Tononi, 2015). The

illustration is a colourised version

of Ernst Mach’s ‘View from the

left eye’.