Perhaps, the most fascinating question about consciousness is the Hard Problem. It’s the problem of explaining why and how subjective experiences arise from complex electrochemical interactions happening in the brain. It is Hard because the working of the brain should be fully described in term of physical interactions, leaving no room for subjective experiences to fit within our current views of the physical world.
This theoretical position is so powerful that scientists often cannot escape from the view that consciousness is an epiphenomenon that plays no role, but just appears as a by-product of information processing in the brain. Even if consciousness researchers may not explicitly admit that they take this position, it is a position difficult to argue against.
The Hard problem is contrasted with Easy Problems, which are about explaining neural mechanisms of objective and observable aspects of consciousness such as the report of conscious perception or voluntary action. Easy Problems are the kind of problems that neuroscientists are usually working on.
The problem of Easy Problems is that they are not easy at all. To work on Easy Problems, we first need to know what kind of functions are associated with consciousness. Many authors have discussed cognitive functions that may require consciousness. The list includes non-reflexive behaviour, intention, imagination, planning, thought, short-term memory, attention, metacognition, emotion and so on. However, there is nothing decisive about such a list at this moment. In fact, empirical studies have revealed quite a range of cognitive functions can be performed outside of conscious awareness even when the stimuli relevant for the task are not perceived consciously. Even functions thought to be tightly linked with consciousness such as attention, working memory, and executive control can be performed in the absence of conscious awareness. Such empirical findings further corroborated the belief aligned with the theoretical cul-de-sac of the Hard problem.
Nevertheless, we should still keep searching for biological functions of consciousness. Easy Problems may not be as sexy as the Hard Problem (because the name suggests they are easy), but the importance of solving Easy Problems should not be undermined. Specifically, we need to understand what is enabled or caused by items that enter consciousness as oppose to those that remained unconscious. This fundamental problem about consciousness is termed the Hard Question by Daniel Dennett, and relatively little attention has been devoted to directly answer this question. Without solving the Hard Question, we are not even ready to attack Easy Problems.
To appreciate the Hard Question, we should distinguish consciousness as a biological phenomenon and conscious experience. When we consider functions of consciousness, they are the functions that are enabled by stimuli that enter consciousness or the functions that can be performed only in awake humans or animals. Functions in this sense should not be confused with the question of what kind of effects conscious experiences (or qualia) exert on physical systems. Because of the conflation of these two different meanings of functions of consciousness, sometimes the validity of asking the Hard Question seems to be questioned. But once this distinction is clear, we are ready to tackle the Hard Question as a target of empirical research.
In our recent article published in Neuroscience of Consciousness, we aimed to offer our tentative answer to the Hard Question by hypothesising possible functions of consciousness. While many cognitive functions are known to occur without consciousness, several lines of empirical evidence in neuroscience and psychology revealed tasks and conditions in which consciousness appears to be necessitated (e.g. trace conditioning). In the paper, we started off by reviewing existing ideas about functions of consciousness and then aimed to propose a unified account for the diverse consciousness demanding tasks in terms of a common underlying function.
The gist of our hypothesis (called the Information Generation Hypothesis) is as follows: the function of consciousness is to generate possibly counterfactual representations of an event or a situation using generative models of the environment and the self, which are learned through interactions with the environment. The term counterfactual here is used to refer to situations or events that are not happening in front of the agent at the present moment. Such representations allow the agent to interact with the information stored from the past (i.e. short-term memory) or to plan ahead through mental simulation of possible future (i.e. imagination or planning). Importantly, this formulation of functions of consciousness explains in a coherent manner a range of non-reflexive cognitive phenomena such as intention, imagination, planning, short-term memory, attention, curiosity and creativity. Another important implication of this hypothesis is that it is a departure from the traditional view that the brain is a passive information processing machine. Instead, the hypothesis suggests that contents of consciousness are the results of active reconstruction of the reality.
Finally, our depiction of the role of consciousness opens up a new avenue for understanding the relationship between consciousness and (general) intelligence. A functional advantage implemented by the information generation architecture enables an agent to perform mental simulations for planning future action sequences for novel goals. This function of consciousness endows the agent with the ability to achieve novel goals that are difficult to attain only with a collection of reflexes. This flexible application of knowledge (i.e. generative/forward models) is a hallmark of intelligent behaviour and hint at the origins of consciousness in the course of evolution.
This article was first published in Oxford University Press