On, of and for the brain

The Evolution of the Sensitive Soul takes readers on a scientific and philosophical journey that gives both an imaginative and incredibly productive perspective on consciousness. The book is an epic poem, albeit in technical but accessible prose, addressing fundamental questions about the relationship between living organisms and the world.

The title of the book comes from Aristotle's three nested levels of the soul: the nutritive and reproductive; the sensitive; and the rational. The sensitive soul 'is the living organisation of the perceiving/feeling animal, which has goals set by its perceptions and feelings that usually, although not always, lead to goal-directed movement' (p.9). This is sentience. Ginsburg and Jablonka see the sensitive soul as specific to animals and as the basis of subjective consciousness, critically not locating this in the rational soul that Aristotle regarded as a human attribute enabling symbolic values such as truth and beauty. For Ginsburg and Jablonka sentience is minimal consciousness.

Minimal consciousness is positioned within a discussion of systems that can self-maintain and reproduce. This is related to autonomy such that a system that has less complexity than its immediate environment is likely to come under direct environmental control. Increased internal complexity facilitates autonomy from this control through off-line processing of incoming data. Systems with this property are referred to as autopoietic, a term developed within the enactivist approach to cognition to capture the sense of reactive embodiment essential to data-processing organisms. It is a profoundly ecological concept that understands life as a response to the environment, as an attempt to maintain coherence across day-to-day and generational time scales. Where enactivist psychologists have tried to relate autopoiesis to traditional cognitive notions of representation, Ginsburg and Jablonka sidestep this potential red herring and delve into biology. If behaviour is the outcome of embodied systems that gain control of the world then the biology of those systems is where the answers lie and we will need to build our theories of cognition from the ground up. This point is made with extensive historical scholarship, as well as exhaustive reference to contemporary findings.

The narrative arc is provided by evolution and the authors travel its length to find a key mark of minimal consciousness and to assay its presence across taxa. What is perhaps most surprising about their pursuit of the distribution problem is that they do not assume that minimal consciousness will be found only within mammals, or even vertebrates. This is because the focus is resolutely upon the contribution of learning and its relation to sentience. The deep dive that the book takes into historical and modern accounts of nervous system evolution and anatomy demonstrates that many of the varieties of learning happen across broad swathes of animal kinds. But also that a good deal of learning can happen under conditions that would not be classified as minimally conscious.

To propose a mark, Ginsburg and Jablonka require a definition of consciousness. This is a task that has exercised scholars across multiple disciplines for centuries. None the less, the authors see reason to be hopeful that a consensus is emerging within neurobiology and from their literature review they draw out what they consider to be the core characteristics that are 'individually necessary and jointly sufficient for consciousness' (p. 98). They are: (1) global activity across the brain; (2) binding and unification of perceptual features; (3) neural selection, learning, plasticity and attention; (4) aboutness (intentionality); (5) the temporal persistence of stimulus effects; (6) experience with a positive or negative valence; and, (7) embodiment, agency and a sense of self.

Unlimited associative learning (UAL) emerges as the candidate mark. UAL 'refers to an animal's ability to ascribe motivational value to a compound stimulus or action pattern and to use it as the basis for future learning' (p.35). UAL can also incorporate novel stimuli, with no previous associations or reflex responses, and it can allow second-, third- and multiple-order associative chains to form thus getting ever further from direct environmental control. UAL delivers organisms that can exploit much broader ecological bandwidths, processing huge amounts of data and using it to update developing hypotheses. Ginsberg and Jablonka honour their debts to Friston and other colleagues who have developed similar ideas within the predictive inference literature. As with these scholars, Ginsburg and Jablonka rely on the emergence of memory systems and the notion of updating past learning in the context of new data. But here effectively cognitive theories are crafted from classical learning theory, and we get to see behavioural science cohering, as it should. To be clear, the claim is not that UAL yields minimal consciousness, but that it requires it. Therefore, where it is found there is a mark of sentience. But this relies upon UAL not occurring in animals that do not meet the seven conditions, something yet to be demonstrated.

The book is big in terms of length but more importantly in terms of intellectual heft. It is designed to survey the landscape but also to direct and to set an agenda. It is not the final word, but rather this book gives a positive hypothesis to be tested. Were the hypothesis to prove false this book will still stand the test of time because its major contribution is a novel and essential synthesis. It is itself a novel compound stimulus with positive valence and multiple associations. It is also an enormous achievement by two gifted scholars.

- Reviewed by Tom Dickins, Professor of Behavioural Science, Middlesex University