daniel_c_dennett's picture
Philosopher; Austin B. Fletcher Professor of Philosophy, Co-Director, Center for Cognitive Studies, Tufts University; Author, From Bacteria to Bach and Back
Philosopher, Tufts University Author, Freedom Evolves

I believe, but cannot yet prove, that acquiring a human language (an oral or sign language) is a necessary precondition for consciousness–in the strong sense of there being a subject, an I, a 'something it is like something to be.' It would follow that non-human animals and pre-linguistic children, although they can be sensitive, alert, responsive to pain and suffering, and cognitively competent in many remarkable ways–including ways that exceed normal adult human competence–are not really conscious (in this strong sense): there is no organized subject (yet) to be the enjoyer or sufferer, no owner of theexperiences as contrasted with a mere cerebral locus of effects.

This assertion is shocking to many people, who fear that it would demote animals and pre-linguistic children from moral protection, but this would not follow. Whose pain is the pain occurring in the newborn infant? There is not yet anybody whose pain it is, but that fact would not license us to inflict painful stimuli on babies or animals any more than we are licensed to abuse the living bodies of people in comas who are definitely not conscious. If selfhood develops gradually, then certain types of events only gradually become experiences, and there will be no sharp line between unconscious pains (if we may call them that) and conscious pains, and both will merit moral attention. (And, of course, the truth of the empirical hypothesis is in any case strictly independent of its ethical implications, whatever they are. Those who shun the hypothesis on purely moral grounds are letting wishful thinking overrule a properly inquisitive scientific attitude. I am happy to give animals and small children "the benefit of the doubt" for moral purposes, but not for scientific purposes. Those who are shocked by my hypothesis should pause, if they can bear it, to notice that it is as just as difficult to prove its denial as its assertion. But it can, I think, be proven eventually. Here's what it will take, one way or the other:

(1) a well-confirmed model of the functional architecture of adult human consciousness, showing how long-distance pathways of re-entrant or reverberant interactions have to be laid down andsustained by the sorts of self-stimulation cascades that entrain language use;

(2) an interpretation of the dynamics of the model that explains why, absent these well-traveled pathways of neural micro habit, there is no functional unity to the nervous system–no unity to distinguish an I from a we (or a multitude) as the candidate subject(s) subserved by that nervous system;

(3) a host of further experimental work demonstrating the importance of what Thomas Metzinger calls the phenomenal model of the intentionality relation (PMIR) in enabling the sorts of experiences we consider central to our own adult consciousness. This work will demonstrate that animal cleverness never requires the abilities thus identified in humans, and that animals are in fact incapable of appreciating many things we normally take for granted as aspects of our conscious experience.

This is an empirical hypothesis, and it could just as well be proven false. It could be proven false by showing that in fact the necessary pathways functionally uniting the relevant brain systems (in the ways I claim are required for consciousness) are already provided in normal infant or fetal development, and are in fact present in, say, all mammalian nervous systems of a certain maturity. I doubt that this is true because it seems clear to me that evolution has already demonstrated that remarkable varieties of adaptive coordination can be accomplished without such hyper-unifying meta-systems, by colonies of social insects, for instance. What is it like to be an ant colony? Nothing, I submit, and I think most would agree intuitively. What is it like to be a brace of oxen? Nothing (even if it is like something to be a single ox). But then we have to take seriously the extent to which animals–not just insect colonies and reptiles, but rabbits, whales, and, yes, bats and chimpanzees–can get by with somewhat disunified brains.

Evolution will not have provided for the further abilities where they were not necessary for members of these species to accomplish the tasks their lives actually pose them. If animals were like the imaginary creatures in the fictions of Beatrix Potter or Walt Disney, they would have to be conscious pretty much the way we are. But animals are more different from us than we usually imagine, enticed as we are by these charming anthropomorphic fictions. We need these abilities to become persons, communicating individuals capable of asking and answering, requesting and forbidding and promising (and lying). But we don't need to be born with these abilities, since normal rearing will entrain the requisite neural dispositions. Human subjectivity, I am proposing, is thus a remarkable byproduct of human language, and no version of it should be extrapolated to any other species by default, any more than we should assume that the rudimentary communication systems of other species have verbs and nouns, prepositions and tenses.

Finally, since there is often misunderstanding on this score, I am not saying that all human consciousness consists in talking to oneself silently, although a great deal of it does. I am saying that the ability to talk to yourself silently, as it develops, also brings along with it the abilities to review, to muse, to rehearse, recollect, and in general engage the contents of events in one's nervous system that would otherwise have their effects in a purely "ballistic" fashion, leaving no memories in their wake, and hence contributing to one's guidance in ways that are well described as unconscious. If a nervous system can come to sustain all these abilities without having language then I am wrong.