The journal, Trends in Cognitive Science, has an interesting paper up: Dimensions of Animal Consciousness. After noting the current consensus that some form of consciousness is present in at least mammals, birds, and cephalopods, it looks at how to evaluate it in various species. The authors take the position that consciousness can be present in varying degrees, but argue that it can’t be assessed on any single scale without missing important dimensions of variation. So, they argue, a multidimensional approach is needed.
Their proposed dimensions are:
P-richness describes the perceptual capabilities of the animal, the level of detail it can perceive in its environment for any specific modality (such as sight, hearing, etc). That richness can be resolved into different components. For example, visual experience depends on bandwidth (how much visual content can be experienced at one time), acuity (the number of differences the animal is sensitive to), and categorization power (the ability to sort perceptual properties into high level categories). These different components can complicate comparisons between species, unless all or most of them are higher or lower.
To probe p-richness rigorously, conscious vs unconscious perception must be distinguished. The paper discusses possible experimental scenarios correlated with consciousness in humans, or lack thereof, such as blindsight, discrimination learning, and other techniques.
E-richness stands for evaluative richness. This is about affects, the conscious feeling of emotions, including valence (feeling good or bad). Similar to p-richness, it can have components, including its own versions of bandwidth and acuity. The authors express doubt about attributing emotions like anxiety or grief to animals, but note that more primal ones like hunger, thirst, or pain are expected to be widespread. (Although they caution against making assumptions without evidence.)
E-richness can be tested with experiments probing motivational trade-offs (such as rats having to decide whether to enter a cold chamber to get at a sugar solution). A trade-off that is crossmodal (somatosensory vs taste for the rats) is stronger evidence for consciousness, requiring crossmodal integration. (Although apparently there is evidence that such integration is not a guaranteed indicator.)
Unity involves integration at a specific point in time. Conscious experience in healthy human adults is highly integrated and unified.
Although there are various pathologies that can undermine that unity, such as split brain patients, who’ve had the connections between their cortical hemispheres separated. Through careful experiments, it can be shown that the two hemispheres no longer communicate with each other, leading to what appear to be two conscious experiences.
Apparently birds are natural split brains. They don’t have the corpus callosum connections between hemisphere that mammals have, implying that a bird may be a couple of conscious entities intimately cooperating with each other. And the nervous systems of cephalopods include a central cerebral ganglia, but local controlling nerve rings for each arm that appear to be at least somewhat independent.
Experiments to test unity in animals are inspired by the experiments on split-brain patients, and involve testing whether information sensed on one side of the body can affect behavior on the other side, or in the case of cephalopods, how much coordination exists between the central senses and the various arms. Other indicators may include whether the animal can have one hemisphere asleep while the other is active, a trait observed in some birds, dolphins, and seals.
Temporality is integration across time, essentially experiencing episodes rather than just snapshots. In its more sophisticated form, the authors argue, this includes mental time travel.
Evidence for this type of capability includes perception of apparent motion, particularly in visual illusions such as the phi illusion (the rapid sequencing of two dots of different color, such that an illusion is generated of one dot moving from one position to another, changing color halfway). It also includes testing for episodic memory and future planning.
Selfhood, self-consciousness, includes awareness of oneself as distinct from the world. In its simplest form, this is awareness of one’s own body. The more sophisticated versions include awareness of one’s own stream of experiences, and mindreading (theory of mind), and at the most sophisticated level, mindreading turned inward. Humans possess all of these. There is evidence that apes and corvids possess some mindreading capabilities, but little evidence that it’s turned inward.
The authors present a table summarizing possible experiments to test and measure these dimensions.
Toward the end, the authors discuss some challenges, one of which is an admission that the dimensions could easily get much more numerous and complicated than what’s presented here. But they cite the need for pragmatism to keep the count down to reasonable number. Another challenge is the need to keep the dimensions distinct from each other, and also to keep them usable for species comparisons.
A few quick observations. The paper struck me as too stingy with the species it sees having episodic memory. I’ve read studies implying it’s more widespread in mammals and birds, albeit with widely varying sophistication. Although maybe more evidence is needed to solidify this conclusion.
I initially thought the authors were too accepting of the mirror test. But they only take it as evidence for self-body awareness rather than self-mind awareness, which is probably the right way to see that test.
Overall, I found a lot to like in this paper. Assessing consciousness across multiple dimensions is, I think, a good approach. It has me wondering if I should find a way to work it into my usual hierarchy, although the goals of that hierarchy are different than this framework.
What do you think of the dimensions? Or the overall approach?