1. Introduction

The debate as to whether or not animals possess consciousness has transcended many thousands of years. The Ancient Greeks thought that animals were organisms with a very different soul from humans (Hurley and Nudds 2006). In the Nicomachean Ethics, Aristotle teaches that animals lack the rational part of the soul, which distinguishes and characterizes man. He proposed that they have only the vegetative and sensitive aspects of the soul, thus experiencing sensations relating only to primary personal needs (i.e. thirst, hunger, sleep, sex, etc.), and lacking the capacity to think in rational terms about such needs and to reflect upon their sensations. Consequently, in Aristotle's view animal behavior is assumed to be generated and driven by current impulses. In addition, animals do not have the concept of themselves as beings that exist through time. In other words, they neither extend their lives with the power of imagination into the future, nor dwell on the past.

The picture of animal life illustrated by the French Philosopher René Descartes was even further devoid of conscious experience. Descartes believed that only man is endowed with a soul and can thereby exist within a spiritual dimension, whereas animals live without awareness in a purely physical universe (Descartes 1637 [1984]). In a word, they are simply automata: they just act and react in mechanical way. Unlike man, who derives from the interaction of two kinds of substances (res cogitans and res extensa) and is, by virtue of this fact, an autonomous agent, animals are mere products of the organization of one substance only (res extensa). For this reason, Descartes claimed in his Discours de la Méthode (1637), their behavior is strictly determined by the laws of Nature. They therefore lead a life which is completely "unconscious".

Although the question of the existence of animal consciousness is long-standing for philosophers, it has a very brief history in science. Charles Darwin (1871) was the first thinker to set the problem within a scientific frame. Before Darwin, the idea that there was a huge and unbridgeable gap between animals and men was common. On the contrary, in Darwin’s view, since both animals and men are to be the products of natural selection, the difference between them is primarily a question of degree (Darwin 1859). Roughly speaking, there are no properties which men have and animals do not have. They all share a common background (the natural world), and the different capacities expressed in the behavior of men and animals are largely determined by the requirements and constraints of the environments within which they develop (Darwin 1859). Darwin’s contribution was to set the problem on the right track, even though we have to recognize that the difference between humans and animals is great, particularly in relation to the astonishing ability shown by human brains in terms of language and the use of symbols.

After Darwin, the issue regarding the nature of animal minds was generally neglected until the end of the last century. Both the development of new techniques for studying the nervous system in order to detect and visualize the brain areas involved in performing certain tasks (i.e. in vivo functional neuroimaging), and the increase of knowledge in the fields of neurophysiology and neuroanatomy, have made it possible to identify the neural correlates of consciousness and better understand the similarities and dissimilarities between humans and animals in terms of conscious experience (Cavanna et al. 2011).

2. Criteria for Defining Consciousness

What level of organization of the nervous system is necessary in order for an animal to possess consciousness? To answer this question, we need a set of criteria which allows us to hypothesize the likelihood that a particular animal is conscious. A useful list of criteria for defining consciousness (Seth, Baars and Edelman 2005) consists of a number of basic brain facts and well-established properties of consciousness.

2.1 Some Basic Brain Facts One approach to determine the presence of consciousness is to compare EEG signatures among species. EEGs of conscious human subjects look very different to those exhibited by subjects who are in unconscious states, such as deep sleep, general anesthesia, epileptic absence seizures, and vegetative states resulting from brain damage (Cavanna and Monaco 2009; Cavanna et al. 2010). We know that consciousness is associated with low-level and irregular activity in the EEG, ranging from 20 Hz to 70 Hz (Berger 1929). Contrastingly, unconscious states are characterized by more regular, slower and high-amplitude waves at less than 4 Hz (Baars, Ramsoy and Laureys 2003). This distinction has been found in all mammalian species studied so far. For example, Cephalopoda can show EEG patterns similar to those of conscious vertebrates (Bullock and Budelmann 1991). Strikingly, similar data have been shown in another invertebrate, the tiny fruit fly (Van Swinderen and Greenspan 2003). Birds which are awake also exhibit EEGs comparable to mammals, although these markedly diverge during sleep (Ayala-Guerrero 1989; Ayala-Guerrero and Vasconcelos-Duenas 1988; Karmanova and Churnosov 1973).

As a second criterion for consciousness, we can assess the activity of the thalamocortical system. Conscious information processing among mammals is associated with specific patterns of interaction between the thalamus and the cortex (Baars, Banks, and Newman 2003). In particular, the lower part of the brain (i.e. the brainstem) is crucial in maintaining arousal, while cortical areas are implicated in the process of creating and manipulating conscious thought (Cavanna et al. 2011). The comparison of anatomical structures among different species implies that the thalamocortical system arose with mammals or mammal-like reptiles more than 100 million years ago; accordingly, consciousness itself might have appeared twice after this divergence, in two reptilian lines which led to birds and mammals (Kardong 1995). This would explain why homologous anatomical structures, which functionally resemble mammalian thalamocortical activity, can also be found in birds (Luo, Ding, and Perkel 2001).

The final criterion relates to the evidence that consciousness implies widespread activity of the brain (Srinivasan, Russell, Edelman, and Tononi 1999). As a result, functional similarities can still be detected between human and animal brains, despite the topological differences in the nervous systems of species whose divergence occurred long ago (Wada, Hagiwara, and Jarvis 2001). For instance, analogies in cerebral functional activity can be inferred by identifying specific molecular markers, such as neurotransmitters, neuropeptides, and neuronal receptors that are present in certain regions of the mammalian brain (Naftolin, Horvath, and Balthazart 2001). Researchers have identified the major neurotransmitters thought to be specific to mammalian brains in the cephalopod nervous system (Hanlon and Messenger 2002).

2.2. The Properties of Consciousness Empirical data suggest that consciousness is characterized by a set of specific properties (Seth, Baars and Edelman 2005). These properties are:

All these properties are hallmarks of human consciousness, but they can be used as criteria in order to determine the existence of conscious states in animals. For example, even the properties that seem to be particular to the human race - such as the reportability of conscious events and the self-attribution of conscious experience - can plausibly be attributed to animals. Of course, animals cannot make accurate verbal reports, however they can give fairly reliable non-verbal indications of their conscious experience by means of indirect methods. Alternative strategies such as the "commentary key" can be used in experiments to allow monkeys to make behavioral comments about what they consciously see or unconsciously guess to see. For instance, monkeys with lesions to half of the visual cortex are able to make discriminations in their blind field by pointing to the correct location of a target or by choosing between colors, but they cannot distinguish between a stimulus in the blind field and a blank display in the intact part of the visual field (Cowey and Stoerig 1995; Stoerig and Cowey 1997). This condition appears to be very similar to human blindsight, in which patients with cortical blindness demonstrate a preserved ability to locate objects, even though they strongly deny that they consciously perceive them. This is further demonstrated by experiments in which rhesus macaque monkeys were trained to press a lever in order to report perceived stimuli in a "binocular rivalry" condition (Logothetis 1998). Binocular rivalry occurs when two different images are presented simultaneously to the eyes, in such a way that one image is seen only by the left eye while the other image is seen only by the right eye. Rather than perceiving both the images superimposed on one another, subjects report consciously seeing one image first and then the other, in an alternating succession.

The properties listed above make consciousness crucial in order to make accurate discriminations between sensory inputs and, consequently, in decision processes. Studies have shown that only conscious sensory information, and not unconscious, activates cerebral regions subserving executive functions (Frackowiak 2004). It therefore appears that consciousness, at least in its basic level of waking state and perceptual awareness, has a key biological function in evolution. The association between consciousness on the one hand, and reproductive behavior and goal-directed survival on the other, is considered by some authors to be indisputable (Baars 2005). This suggests that consciousness can play a very important role in the continuity of animal life.

3. Consciousness in Animals

Demonstrations of consciousness in animal species may require behaviors which satisfy the criteria listed above and are unlikely to reflect simple conditioned responses.

Many animals discriminate between sensory inputs and flexibly alter behavior. William James (1910) suggested that the recognition and integration of the "sense of sameness is the very keel and backbone of consciousness" (p. 240). Pigeons appear to possess some sense of sameness, and therefore maybe consciousness, given that they can discriminate between visual stimuli. Cook and colleagues report that pigeons can identify the odd area or item across a wide range of visual stimuli. For example, Cook, Kelly, and Katz (2003) showed pigeons either a sequence of same (aaaaaa) or alternating (ababab) photographic stimuli. Pigeons showed that they could discriminate these two types of sequences by exhibiting reward-reinforced differences in pecking rate depending on whether they were shown identical or nonidentical pictures. Following acquisition, the pigeons showed successful transfer of this behaviour to novel stimuli. Pigeons can therefore make a same–different discrimination based on only two different items, and can detect stimulus change quite quickly.

Other good examples of behavioural adaption based on sensory input are tool use and observational learning (Masson and McCarthy, 1999; Shanor and Kanwal, 2009). Tool use has been demonstrated in elephants (Chevalier-Skolnikoff and Liska, 1993) and some types of birds, which include carrion crows in Japan utilizing the movements of cars at traffic lights to crack nuts. Primates are also well known tool users. Orangutans use leaves as gloves for holding prickly fruit and pitcher plants as cups, while chimpanzees have been observed carrying sharp sticks as weapons. Primates in captivity including orangutans readily imitate human behaviors accurately, and octopi can learn skills through observation (Edelman, Baars and Seth, 2005).

Animals may also demonstrate conscious thought through the ability to formulate plans to solve problems, implying they have an understanding of their place within the world and within time. Simple examples of such behaviors include food caching. However, more complex planning abilities were shown by a chimpanzee in a zoo in Sweden, who planned ahead to throw rocks at annoying visitors (Osvath, 2009). The complexity of such behavior seems to imply conscious intention, rather than actions which depend on simple stimulus-response associations. Furthermore, while many animals, including octopus and cuttlefish, can exhibit evidence of short and long term memory, is it possible that this primate’s behavior indicates the use of prospective memory, or mental simulation of future events?

An abstract representation of future events may also be exemplified by the behavior of the parrot Alex (Pepperberg, 2008). After witnessing a nut being obscured by a cover and then a food pellet being revealed, rather than simply eating the food pellet, Alex exhibited aggressive behaviors, implying annoyance at the object switch. Alex remembered that the nut was previously in that place, even despite it being invisible to him for a period of time. Furthermore, Alex’s reaction seems to imply that he had formed expectations that the nut would still be present when the cover was removed, i.e. an understanding of object permanence. Studies currently debate the age at which this ability develops in human infancy.

Although animals cannot use language to communicate with humans, they may demonstrate a degree of language use or understanding which could reflect conscious thought. Studies of prairie dogs (e.g. Slobodchikoff et al., 1991; Slobodchikoff, 1998) indicate these animals use specific calls to describe different predators. These calls therefore possess true meaning and have clear adaptive survival value. Closer study of prairie dog language indicates the use of modifiers: sounds which reflect differences in attributes such as colour and size. Such language demonstrates evidence of awareness of changing environmental conditions and perhaps even a desire to communicate such changes. In relation to understanding human language, one border collie called Rico has a vocabulary of more than 200 words. Experiments showed that Rico correctly retrieved 37 of 40 randomly chosen toys by name from his collection of 200 toys. But more striking is this dog’s ability to learn the names of new toys. When researchers placed a new toy amongst his familiar toys and asked him to get the toy using a word he had never heard before, Rico retrieved the new toy 70% of the time. This could provide evidence of conscious thought and even deductive reasoning, whereby Rico realised that the word must relate to the unknown item as it is was not the same sound associated with the other items he knows (Stein, 2004). There was also evidence that Rico added these new toy names to his vocabulary without further instruction, learning new words in a ‘fast-mapping’ fashion (Kaminski et al., 2004), which is similar to the way young children aquire new words.

Some animals such as parrots and primates including chimpanzees, gorillas and bonobos, have been taught to communicate in human language using picture boards or sign language. Studies of these animals illustrate key evidence of conscious awareness, as reflected in the animal’s ability to describe objects in their environment. Critically, these animals can combine words or sounds they have previously been taught to create a new word to describe a novel object. For example, the parrot Alex (Pepperberg, 2008) described an apple as a ban-erry (combining part of the words banana and cherry) and cake as ‘yummy bread’. On the first occasion that Kanzi the bonobo ate kale, she described it as ‘slow-lettuce’, presumably because it took longer to chew and swallow than the normal lettuce she was familiar with (Savage-Rumbaugh and Lewin,1994). These combinations of new sounds and words can’t reflect simple stimulus-response associations and therefore provide relatively strong evidence of accurate report.

An aspect of human consciousness is a personal or internal awareness of the self and reflection on one’s own conscious experience, involving elements such as metacognition. Mirror-self recognition (Edelman and Seth, 2009; Shanor and Kanwal, 2009) has thought to have been demonstrated in chimpanzees (Povinelli et al., 2003), orangutans (Suarez and Gallup, 1981), bonobos (Hyatt and Hopkins 1994; Walraven et al. 1995), possibly gorillas (Patterson and Cohn, 1994), rhesus monkeys (Rajala et al., 2010), elephants (Plotnik et al. 2006), dolphins (Reiss and Marino 2001), magpies (Prior et al. 2008) and, more controversially, pigeons (Epstein et al., 1981). For instance, experiments have verified with quantitative and videographic evidence that rhesus monkeys with surgical implants in their heads use mirrors to inspect their implants, as well as other parts of their bodies which they are not able to see (Rajala et al. 2010). Fewer experiments have attempted to investigate metacognition in non-human species, but reports suggest that parrots may show evidence of insight into changes in their own knowledge by discriminating a discrimination (being aware of distinguishing changes: Pepperberg, 2008).

Likewise, dolphins have been shown to be capable of complex intellectual and cognitive activities that some believe are equal or superior to non-human primates (Marino, 2002; Marino et al. 2007). Further, they appear to have a sense of self-image including an ownership of individual body parts (Herman et al. 2001). Dolphins can be trained in the use of gestures which they can learn to associate with specific body parts, and can move, shake or indicate specific body parts when asked, such as by using a specific region of the body to touch a specific object. They will also repeat specific behavior when asked to do so. Moreover, they can mimic not just other dolphins, but humans, and can mirror human actions, such as waving their pectoral fin when a human waves an arm, raising their tails when observing a human raising their leg (Herman 2002). These results suggest that dolphins have self-consciousness and are conscious of their actions as having agency ad ownership (Mercado et al. 1998, 1999; Reiss and Marino 2001).

If animals can understand themselves as individual entities, can they appreciate the existence of other organisms as entities with similar functions? Some researchers argue that certain animals appear to exhibit appreciation of other animals’ abstract mental states, or a rudimentary ‘Theory of Mind’. Such awareness may be implied by the deceptive behavior of ravens, who re-hide food after conspecifics observe their actions, but not if other birds appear to ignore them (Bugnyar and Heinrich, 2005). When a dog owner points at something, his pet seems to understand he should look not towards the finger, but towards the direction at which it is pointed (Horowitz, 2009). Could this express awareness of his owner’s intent? Jackdaws can also use human eye gaze and pointing as cues to behavior suggesting they have some rudimentary understanding of human desire or intent (Shanor and Kanwal, 2009).

Other experiments showed that dominant and subordinate apes competed with each other for food. These findings led researchers to conclude that "at least in some situations chimpanzees know what conspecifics do and do not see and, furthermore, that they use this knowledge to formulate their behavioral strategies in food competition situations" (Hare et al. 2000). It is worth noting here that the theory of mind debate finds its origin in the hypothesis that primate intelligence on the one hand, and human intelligence on the other, are specifically adapted for social cognition (Byrne and Whiten 1988). As a result, it has been argued that evidence for the ability to attribute mental states in a wide range of species might be better sought in natural activities rather than in lab experiments which place the animals in artificial situations (Allen and Bekoff 1997; Hare et al. 2001; Hare and Wrangham 2002).

Perhaps some of the strongest evidence of the likelihood of animals possessing conscious thought are observed behaviors whereby animals act against natural survival responses (Masson and McCarthy, 1999). For example, one report described an elephant in Kenya attempting to rescue a young rhino trapped in mud despite the rhino’s mother’s aggressive response. In another specific case, when a bull elephant was shot, two younger elephants attempted to rescue and lead him away (Denis, 1963; cf. Carrington, 1958). It has also been observed that dolphins and whales support conspecifics if they need help to breathe (Bearzi and Stanford, 2008), and move injured companions away from attack and danger If we are to accept that such seemingly altruistic responses are unlikely to reflect simple survival instincts, we may have to consider the possibility that they imply some form of conscious will.

4. Conclusions

Investigations of consciousness in animals encompass examination of both the structure and function of the central nervous system and observable behavior. Studies investigating the similarity of EEG patterns, neural function and markers of neuronal activity often indicate a degree of overlap amongst humans and other animals, while observational studies have documented a range of behaviors demonstrated by many species which cannot be ascribed to basic stimulus-response associations and are more likely to reflect conscious processing. Our judgment of whether animals possess consciousness will inevitably depend on the criteria used. However, examination of available behavioral evidence suggests a possible hierarchical organization of consciousness within the animal kingdom (Figure 1).

Criteria which can be used to determine levels of consciousness include level of sensory awareness, degree of behavioral flexibility, complexity of communicative abilities and social interaction, and the presence of higher order abilities such as self recognition, mental reflection and theory of mind (Table 1).

Given the diversity of life in the animal kingdom, it is clear that the question of whether animals possess consciousness is far too simple. Rather, we should ask: at what level does a particular species demonstrate conscious awareness? The challenge for future research will be to conduct investigations which overcome species barriers in order to provide more objective scientific evidence of the conscious abilities easily ascribed to animals based on observed behavior.