- seCtIon one: tHe PRoBLem
opened things up further. But it was nearly three decades before the sudden
explosion of interest in consciousness in the 1990s.
From around the 1950s to the 1990s the ‘first-generation’ cognitive sciences had
conceived of the mind in terms of abstract, language-like representations (Lakoff
and Johnson, 1999, pp. 77–78) and relied heavily on analogies with digital com-
puters, but increasingly people began to think more in terms of interconnected
networks that change over time. From this connectionist approach came the idea
of the neural network, which revolutionised the study and creation of artificial
intelligence (see Chapter 12). This movement, along with the embodied philoso-
phy of Merleau-Ponty, contributed to the emergence of the ‘second generation’ of
cognitive science, which recognised that brains are always found in bodies, and
bodies in environments – both physical and social.
And thinking about cognition as ‘4E’ (Menary, 2010) – as embodied, enactive,
embedded, and extended (involving other objects and people in the environ-
ment) – opens up much more space for experience than does a computational
brand of cognitivism. As the authors of The Embodied Mind put it: in the embodied
paradigm, ‘cognition and consciousness – especially self-consciousness – belong
together in the same domain. Cognitivism runs directly counter to this conviction
[. . .] for cognitivists, cognition and intentionality (representation) are the insepa-
rable pair, not cognition and consciousness’ (Varela et al., 1991, p. 173). Thinking
about the kinds of experience that come from having a body with particular
sensory and motor capacities, and from the feedback between these capacities
and the environment, gives us an alternative to trying to uncover consciousness
through the neurons alone.
An approach that combines the insistence on feedback between brain, body,
and world with a firm basis in brain function is that of predictive processing, the
idea that brains are essentially prediction machines, constantly trying to match
incoming sensory inputs with their own expectations or predictions (Clark, 2013).
This is a modern version of Helmholtz’s idea of ‘unconscious inference’ and British
psychologist Richard Gregory’s (1966/1997) much later notion that perceptions
are guesses, or hypotheses, about the world. The difference is that with advances
in neuroscience and computation we can begin to work out how the embodied
brain builds its predictions and adapts the body’s responses to the world.
Dynamics-based paradigms like this give us challenging new ways of thinking
about the status of the brain in relation to the mind and consciousness, which we
will return to in Chapter 3. In principle, they also link to other theories that stress
the contexts of consciousness, such as social constructionism, a movement which
built on the developmental psychology of the Soviet psychologist Lev Vygotsky
in the 1930s, and which investigates how reality as we know it is constructed
through social interactions. They have certainly begun to generate resonances
within cognitive approaches to the study of literature (Caracciolo and Kukkonen,
2014). But in practice, the methods and questions of 4E and predictive processing
have not yet tended to expand very far into the study of language, history, and
culture.
Even now, after centuries of philosophical and psychological inquiry, our under-
standing of how behaviour and ‘introspection’ relate to consciousness leaves a lot
to be desired (Costall, 2006), and even a definition of consciousness that everyone
‘The mind is not in the
head’
(Varela, 1999, p. 72)