1. Introduction. The Temptation of Epiphenomenalism

One of the most enduring and intriguing questions for both philosophical and scientific researchers is whether we have conscious minds capable to control and produce the motivations for all our actions. In the light of our common experience, an affirmative answer to that question would be but a platitude. Indeed, brains seem to be capable to create a great variety of mental events. Love, hate, sadness, joy, sorrow, pleasure, shame, grief, delight, and resentment are only a few of the many different psychological states composing our rich mental lives. However, on more accurate reflection the solution to the problem of the nature of consciousness would not appear as evident as it may seem at first sight.

There are, in fact, philosophers and neuroscientists who firmly believe that mental properties, particularly the conscious ones, are wholly epiphenomenal with respect to brain processes (Edelman 2004, 2007; Fuster 2003). According to epiphenomenalism, mental properties are superfluous by-products of the function of our cerebral mechanism, just like the images which are reflected by mirrors are not made by glass, and the shadows which objects cast on the ground are not parts of those objects. This view is not new: in a famous conference held at the British Association for the Advancement of Science, Thomas Henry Huxley compared consciousness to the steam whistle of a locomotive (Huxley 1884). In contrast with Descartes, Huxley did not consider animals as unconscious machines, but was very perplexed with regard to the exact function of consciousness and hypothesized that conscious states played no role in behavioral mechanisms (Huxley 1874). Just as the steam whistle of a locomotive did not influence the work of the locomotive's motor, he thought, so animal consciousness could neither cause nor modify animal behavior. Being also a strenuous advocate of Darwin's theory of natural selection, Huxley assumed for reasons of biological continuity that there are no differences between animal and human consciousness (Huxley 1884).

We can affirm that the modern shape of the problem of epiphenomenalism was set with Huxley, even though he never used this word in his writings. In effect, the Modern Age (post-Cartesian) thinkers did not tend to contrast sharply the concepts of mind and body. Ancient Greeks had a much broader idea of mind, closely linked with bodily functions (Bremmer 1983). Mind (i.e. the soul) was considered the principle of life capable to animate the body in order for it to perform its basic biological processes, such as breathing, digestion, procreation, growth, motion, and, for humans, also other sophisticated processes of life, such as thinking, perceiving, imagining, and reasoning. Of course every school of ancient Greek philosophy had its own concept of "soul”. For instance, Aristotle thought of the soul as the body's system of active abilities to accomplish the vital functions that organisms naturally perform, e.g. nutrition, movement or thought (Nussbaum and Rorty 1992). On the other hand, Stoic philosophy of mind conceived the soul itself as a corporeal entity (Inwood 2003). This position was similar to that of Epicurus, who taught the soul to be a kind of body composed of atomic particles (Kerferd 1971). Perhaps Plato and the Pythagoreans held the closest concept of soul to the Cartesian view. They maintained it to be as something incorporeal and able to exist independently of the body ( Lorenz 2008; Huffman 2009). Nevertheless, it is only after Descartes' philosophy that the debate of mental causation was to be set down in its modern form. An important echo of this debate is to be found in the discussion upon the automatism of behavior raised by Huxley in the second half of the nineteenth century.

Positions similar to those embraced by Huxley are still held by some contemporary philosophers and neuroscientists. Epiphenomenalists would be willing to explain the origin of consciousness in the same way as we can physically explain how mirrors produce reflected images or bodies cast their shades on the ground. According to this explanatory model, all our psychological states should, in theory, be accounted for entirely in terms of scientific vocabularies which contain no mental concepts.

A similar approach, at least in relation to its practical consequences for psychological research, was maintained by Burrhus Skinner with his theory of "radical behaviorism”. Skinner supported the view that mental terms could be completely paraphrased in behavioral terms, or eliminated from explanatory discourse altogether (Skinner 1974). Therefore, all accounts of human behavior would have to be given in neutral and objective terms, such as stimulus, response, conditioning, reinforcement, and so on. This position has some analogies with epiphenomenalism in that it considers consciousness a nonphysical entity which has nothing to do with behavior.

Such epiphenomenalist lines of reasoning imply that a rigorous discourse on human actions should deny the reality of consciousness, and thereby of all mental states correlated with this phenomenon. In fact, when human beings express propositions about conscious states, they actually intend to speak of other things, specifically of certain brain physical states which are to be the unique causes of all their bodily dispositions. In William James' words, "consciousness … would appear to be related to the mechanism of [the] body simply as a collateral product of its working, and to be completely without any power of modifying that working” (James 1890). In a sense, arguing that conscious mental phenomena are causally real would be like arguing that the black spots which are on the tails of some fishes are real eyes capable of seeing and not just evolutionary tricks for misleading predators.

The solution offered by epiphenomenalism to the problem of the nature of consciousness can be roughly summarized in the assertion that such a problem does not exist because we have no consciousness, but only the illusion of having it. Although epiphenomenalism could be somehow attractive, it does not present a satisfactory solution to the problem of what consciousness really is. The present paper will show that the epiphenomenalist perspective does not offer a consistent account of certain neuropsychological phenomena which seem to be intrinsically subjective.

We will start by examining the arguments that drive some philosophers and neuroscientists to regard consciousness as an empty concept involving outdated thinking.

2. Causal Links

Epiphenomenalism upholds that the conscious mind is not part of the physical world. This implies that, given the physical causal closure of the universe, conscious mental events cannot interact with the physical reality in any way. If we think of ourselves as consciously acting agents, the epiphenomenalist claim sounds counterintuitive: is it plausible that our conscious will cannot influence the physical world?

In order to answer this question we first need a better understanding of exactly what it means to claim that conscious mental events cannot influence physical events in any way. This doctrine implies that the only possible type of causation we have to deal with is the so-called bottom-up causality, i.e. the causation that goes from the physical level to the conscious one. According to this concept of causality, the irresistible desire for an apple pie cannot actually cause the act of eating a slice of pie, since the account of our behavior is to be determined only at the physical level, where physical entities move other physical entities. However, if all causation processes belong to the physical level, how could our ontological catalogue list phenomena which are not included in causal accounts of behavioral expressions given in physical terms only? A reasonable philosophical precept (epitomized by Occam's razor) warns us that ontology should not be expanded without necessity. Moreover, there is the problem of defining the nature of those phenomena. If these entities were completely different from physical processes, then conscious mental events would necessarily belong to a distinct ontological domain, but it is easy to conclude that in such a case epiphenomenalism would be just like a spurious kind of dualism. On the other hand, if conscious mental events were physical phenomena of a very special nature, how could we distinguish the physical events capable to cause other physical events from the ones which have no causal power at all?

We cannot actually make a distinction of this kind by means of the third person vocabulary that scientists generally use to depict their objective vision of the world. Therefore, those who trust epiphenomenalism would have to include in their ontological catalogue states, events, and processes susceptible to be described exclusively in terms of the first person perspective, and to identify these states, events, and processes with non-causal physical phenomena. In addition, this sort of phenomena would have to be put together with all the other states, events, and processes liable both to be described in terms of the third person perspective and to be identified with causal physical phenomena. However, it is by no means clear why some processes – whose nature is basically physical – would have to be described exclusively in terms of the first person perspective rather than third person perspective.

In addition to these conceptual difficulties, a further grave quandary is whether we maintain the division between causal and non-causal physical events. In other words, taken for granted such a division, what in this case would the principle of the physical causal closure precisely mean? The principle of the physical causal closure states that if a physical event has a cause that occurs at t, it has a physical cause that occurs at t. Jaegwon Kim (2005) correctly observes that the "physical causal closure does not by itself exclude nonphysical causes, or causal explanations, of physical events.” For instance, there could be a nonphysical causal explanation of a physical event being the first ring of a chain of other numerous physical events. According to Kim (2005), in order to rule out this kind of explanation, we need an exclusion principle such as the following: if an event e has a sufficient cause c at t, no event at t distinct from c can be a cause of e.

Following the principle of exclusion, the sufficient cause c of e at time t may be either physical or mental. However, this instance is ruled out if the principle of exclusion is linked to the principle of the physical causal closure and the event e is identified with an event whose nature is purely physical. As a result, the two principles joined together hold that only physical events can cause other physical events.

It is important to highlight that these two principles are not stricto sensu in contrast with the folk psychology view that there is a conscious mind in every human being. Moreover, these assumptions are consistent with the hypothesis of psycho-physical parallelism, according to which there could be a distinct domain of specific conscious mental events coming to occur whenever other particular physical events come to occur. Still it is unconditionally denied that conscious mental phenomena can causally interact with physical phenomena. In fact, in order to be closed, any causal explanation is to be expressed as a chain of purely physical events. This leads us back to our previous question: if we accept that both epiphenomenal events and non-causal physical entities can by no means be part of scientific accounts given in causal physical terms, then why should we expand our ontology without necessity by including allegedly redundant conscious phenomena?

3. An Argument from Quantum Physics?

An interesting argument against the plausibility of epiphenomenalism can be derived from a specific interpretation of the theory of quantum mechanics. It is well-known that the act of measurement is a crucial aspect from the perspective of quantum theory. The implications and the account of this process has been the subject of controversy for more than seven decades and the debate does not seem to be closed yet.

The fundamental point of the argument from quantum physics holds that the observer's consciousness plays a key role in the collapse of the wave function to a certain state, described by a second-order differential equation by Erwin Schrödinger. The root of this idea can be found in the so-called Copenhagen interpretation of quantum mechanics. Although the Copenhagen interpretation is not a homogenous view (Howard 2004), Heisenberg (1955) appears to be the one who first coined the term and developed the underlying philosophy as a unitary interpretation (Heisenberg 1958). On the other hand, Niels Bohr – the Danish physicist commonly regarded as the father of the Copenhagen interpretation – never seems to have emphasized or privileged the role of the observer in the wave packet collapse (Howard 1994). Bohr argued for an interpretation of complementarity with regard to the wave-particle duality which is incompatible with Heisenberg's interpretation of wave function collapse (Gomatam 2007). Bohr's view regarding the wave function was more moderate than Heisenberg's and based on epistemological concerns, rather than ontological commitments. When Bohr referred to the subjective character of quantum phenomena he was not referring to the conscious intervention of the observer in the process of measurement, but to the context-dependent status of all physical observations (Murdoch 1987; Faye 1991). However the drastic theoretical move – outlined in Heisenberg's writings – that quantum measurement is to be understood by involving the observer's act in addition to a physical process, has become the core of the so-called Copenhagen interpretation.

This idea was further developed by other physicists. Von Neumann (1932) postulated an ad hoc intervention of an observing system in order to account for the collapse or reduction of the wave function through measurement. Quite cautiously, he never claimed that the observing system had to be conscious (von Neumann 1932). In contrast to von Neumann's view, London and Bauer (1932) attributed to the oberver's consciousness only the key role in understanding the process of quantum measurement. Such a proposal was later expanded on by the physicist Eugene Wigner, according to whom "it was not possible to formulate the laws (of quantum theory) in a fully consistent way without reference to consciousness” (Wigner 1967).

The proposal to consider consciousness as causally involved in physical state reductions was further developed following Heisenberg, von Neumann, and Wigner (Stapp 1993, 1999, 2006; Schwartz et al. 2005). Undoubtedly, this approach challenges the epiphenomenal position. In fact, how is it possible for consciousness to be non-causal if it can bring about the collapse of the wave packet? If those who champion the Copenhagen interpretation are right, then epiphenomenalism should be completely refuted. On the one hand, this approach gives a fundamental causal power to consciousness in understanding the universe; on the other hand, it has the serious shortcoming of putting consciousness outside the physical world. In fact, if consciousness really causes the collapse of the wave function, then it must be a process that is not be describable by Schrödinger equation, because otherwise it would be caught in an infinite regress. Based on these arguments, consciousness should be a nonphysical entity. Therefore, if both the principle of the physical causal closure and the principle of exclusion discussed in the previous section are true, the so-called Copenhagen interpretation of quantum mechanics does not provide a strong argument for refuting epiphenomenalism.

4. Consciousness and Causality

In view of the foregoing reflections, we would have to be very reluctant to claim the distinction between causal and non-causal physical events or between causally efficacious physical and causally efficacious nonphysical events, although these distinction are of course logically possible. Accordingly, both philosophers and neuroscientists well-disposed to epiphenomenalism have to maintain that the nature of conscious mental processes cannot be physical, since, by definition, a state/event/process exists in physical terms only if it has the property of influencing and exerting causal effects over other physical entities.

An interesting type of neuroscientific approach is exemplified by the epiphenomenalist position held by Gerald Edelman, a leading neuroscientist who has given remarkable contributions to the study of mind and consciousness and their place in nature (Edelman 1989, 1993, 2007). According to Edelman's approach ("Neural Darwinism”), different configurations or patterns of neurons compete with each other to gain constancy and stability within the brain. The Neural Darwinism approach holds that groups of neurons and the neural patterns and configurations which nerve cells form ("neural networks”) are subject to natural selection, just like biological species are evolutionarily selected by the environment. Specifically, Edelman's theory of "neuronal group selection" postulates that anatomical connectivity in the brain occurs via selective mechanochemical events that take place epigenetically during development. This process creates a structurally diverse primary repertoire by differential reproduction. A second selective process occurs during postnatal behavioral experience through epigenetic modifications in the strength of synaptic connections between neuronal groups, thus creating a diverse secondary repertoire by differential amplification.

In Edelman's view, human consciousness depends on and arises from the uniquely complex physiology of the human brain. He advanced a theory of how the brain generates different levels of consciousness through multiple parallel re-entrant connections between individual cells and between larger neuronal groups, in which he endorsed an epiphenomenalist position with regard to consciousness, which is central to the scope of this paper (Edelman 2004, 2007).

It has been argued that the question whether consciousness can have a causal role in determining behavior and other mental states should find an answer supported by both conceptual and empirical considerations. (Flanagan 1992; Heil and Mele 1993; Searle 2004). Contrary to this view, Edelman seems to give pre-eminence to theoretical arguments over empirical results. In fact his thesis – that consciousness is not causal – is almost exclusively based upon the following theoretical argument:

Edelman's theory is central to the discourse on epiphenomenalism advanced so far. In fact, the phenomenal properties which he refers to are necessarily implied by the underlying neural activity of the brain. Strictly speaking, those phenomenal properties are not redundant but absolutely non-causal, even though they have a sort of physical nature. In addition, those properties have to be seen as by-products, since they cannot play any specific role in our scientific account of natural phenomena. Therefore, neither the principle of physical causal closure nor the principle of physical exclusion seem to be violated by Edelman's perspective.

In our view, the arguments put forward by Edelman in order to demonstrate the epiphenomenalist nature of consciousness raise a number of issues. What is most unclear is the very nature of the necessary relationship between the causal physical events (i.e., the cerebral processes or C') and the non-causal physical events (i.e., the mental processes or C). If we accept that this relationship is necessary, there is no reason to assume an ontology in which conscious mental processes and physical processes within the nervous system are distinct. If a certain property is necessarily implied by certain physical processes (in such a way that the latter could not bring about the same effect without the former, as Edelman claims), then either that very property and those physical processes are different aspects of the same entity, or that very property is part of the co-occurring physical processes. From the logical point of view, an effect cannot find its cause in one event which is the result of the sum of a causal physical state and a non-causal physical state, since the non-causal physical process cannot play any causal role at all (Heil and Mele 1993). In fact, what Edelman believes to be non-causal, "the phenomenal transform”, must be provided with causal powers.

In this sense, Edelman's theory with regard to the epiphenomenalist nature of consciousness appears to be anomalous. A "true” epiphenomenalist would, in fact, plausibly think of the causal relationship between physical and conscious states as contingent rather than necessary. Thus, epiphenomenalists should be willing to accept the zombie argument, since it is logically possible to accept that, if conscious mental processes are contingent, there could be possible worlds in which they do not bring about any behavioral effect.

In addition to these theoretical considerations, empirical data can raise other reservations with regard to epiphenomenalism. For example, if we agree on depriving mental entities of their causal role, we would encounter difficulties accounting for a host of well described neurological conditions. These include, but are not limited to, blindsight (cortical blindness with preserved ability to locate objects), unilateral neglect syndrome (loss of ability to detect information coming from the left side of the body), allochiria (experience of a sensory stimulation at the contralateral side to the applied stimulus), anosognosia (denial of gross neurological deficit), prosopognosia (inability to recognize familiar faces), and somatoparaphrenia (a condition in which patients deny ownership of a limb or an entire side of their body). Arguably, these neurological disorders can be explained at least to some extent in terms of a dysfunction in the causal role played by consciousness in dealing with perceptive or proprioceptive information.

The understanding of somatoparaphrenia is an exemplar case, based on the concept of verbal manifestations commonly referred to as propositional attitudes in the tradition of analytic philosophy (Bisiach and Geminiani 1991). Propositional attitudes are all the expressions whose contents consist of subjective beliefs, desires, intentions, fears, etc. In case of a patient showing somatoparaphrenic symptoms, it is plausible to suppose that a dysfunction in the conscious processing of proprioceptive information about the patient's limb (for instance, the left leg), results in the patient holding the belief that the leg does not belong to his body. Patients with somatoparaphrenia will therefore verbally deny that they own that leg, and some of them have in fact been reported trying to reject the limb that they perceive as alien (Critchley 1974).

Somatoparaphrenia has been described, with a few exceptions, in patients suffered from right parietal (or parieto-occipital) lobe injury – and almost invariably concerns the left side of the body. This condition is usually associated with motor and somatosensory deficits, and with the syndrome of unilateral spatial neglect. In a study on 79 acute stroke right-brain-damaged patients (Baier and Karnath 2008), 12 patients showed anosognosia for hemiplegia. Eleven out of these 12 patients exhibited somatoparaphrenic symptoms, and 6 among them displayed the strong belief that their limbs belonged to another person. In other cases, body parts can be just felt by the patient as separated from the body (Starkstein et al. 1990). More complex symptoms have been described: for instance, a patient can refer to the affected limb as "a make-believe leg” (Levine et al. 1991), or as "a baby in bed” (Richardson 1992).

The spectrum of somatoparaphrenic symptoms is wide, however a distinction can be drawn between misidentifications that can be corrected by patients when the error is pointed out by the examiner, and delusions that stubbornly resist to the examiner's demonstration (Feinberg et al. 2005). It is referred to fully-fledged somatoparaphrenia only in the second category of symptoms.

In sum, somatoparaphrenic phenomena do not imply a mental illness and can be characterized as follows (Vallar and Ronchi 2009):

Overall, available data suggest that patients showing somatoparaphrenic symptoms suffer from impairments in the higher-level processes concerned with body awareness and ownership. Therefore it seems reasonable to hypothesize that if certain behaviors do not occur without specific conscious sensations accompanied by the beliefs which refer to them (i.e. mental representations of the body), then those specific conscious sensations and their consequent beliefs must play an important causal role in the process that produces this kind of behavior.

5. Conclusion

In specific scientific contexts, it seems mandatory to apply concepts which carry a commitment for a causal role for consciousness. For the sake of the unity of science, it seems justified to take the physical causal closure for granted; on the other hand, it does not seem as well justified to take for granted the clear-cut distinction traced by epiphenomenalism between the physical world and the conscious mind. However, neither the common version of epiphenomenalism (in which conscious states are contingent), nor its variant proposed by Edelman (in which consciousness and the physical world are necessarily intertwined), appear to be a valid theory to explain the nature of consciousness. The theoretical and empirical arguments advanced in this article show that it is likely for consciousness to play fundamental roles in the genesis of behavior. Undoubtedly much more is to be done, especially on the side of the empirical research, in order to unravel the actual brain mechanisms of conscious causal processes.