Posts Tagged ‘turing test’

No sentience without sapience

June 26, 2014

There was a great deal of publicity last week but I am not very convinced that the computer program Eugene Goostman actually passed the Turing test. But whether it did or not, I got to wondering how to distinguish sapience from sentience.

I find that I tend to use “sapience” to imply the capability for thought while I take “sentience” to be a quality of consciousness of self. Which of course leaves rather diffuse and undefined what precisely “thought” involves and what “consciousness of self” consists of. But is sapience linked to sentience? Can one have one without the other? Or does the quality of being conscious only become possible once thought exists?

Rene Descartes’ Cogito ergo sum (I think therefore I am) should perhaps be modified to be Cogito ergo, ego ut sit (I think therefore I may be). “I think therefore I am” requires a pre-conception of individuality, of the “I”. To be aware of the “I”, to be conscious of oneself and to be able to articulate that consciousness would suggest that thought is already present before consciousness of self can come into play. Clearly my computer “thinks” in a fashion, as do many animals – in their fashion. But while some minimum capability for thought may be necessary for consciousness, it is also clearly not sufficient. A certain level of sapience may be necessary for sentience but sapience does not necessarily lead to sentience. Some other attribute or quality is required for a thinking entity to be said to have the level of consciousness necessary for sentience.

The Turing test is, I think, a test of reaching a particular level of sapience but it is not a test of sentience. But I also think that there is a scale of sapience. All  “artifical intelligences” show varying levels of applying “thought” and could be said to be sapient to some degree. Sapience would seem therefore to be on a continuous scale. Many animals and birds also exhibit some level of thought and clearly exhibit different degrees of sapience. But chimpanzees and gorillas and dolphins and even elephants seem to recognise themselves in a mirror while monkeys do not. They would seem to have different levels of self-consciousness and – it would seem – different levels of sentience. I take gorillas and chimpanzees and maybe elephants to be sentient – just – but not dogs or cats. Is there then a scale of sentience which is constrained (or enabled) by, and depends upon, an entity’s position on a scale of sapience?  I suspect that whatever it is I intuitively consider to be sentient depends upon a combination of sapience and the level of consciousness of self of an entity.

Therefore my tentative definitions / conclusions become

  1. Entities may be “alive” or “inert”.
  2. Only some entities are sapient to any significant degree but sapience is independent of being alive.
  3. There is no sentience without sapience.
  4. Only some “living”, sapient entities are sentient.
  5. Sentience is a composite quality and – I propose – depends on the level of sapience and the level of consciousness exhibited by an entity.

 

sapience and sentience

sapience and sentience

Advertisements

%d bloggers like this: