We talk a lot about "Artificial Intelligence". Who would you turn to as an expert of "Real Intelligence" - especially in non-human animals? Or could you recommend any books / articles on the topic of defining / characterising intelligence?

(There must be a lot about #IQ and IQ tests but apparently that's not a great measure so.. What is?)

#AI #RealIntelligence #ArtificialIntelligence

@elduvelle_neuro I'm interested to see the answers you get, but my suspicion is that there is no good answer.
@neuralreckoning @elduvelle_neuro I've been very inspired by Jacob Foster and Erica Cartmill and the crew at DISI.org. They have a truly wonderful community of researchers that they bring together for a summer institute, a podcast, and now a dedicated center at IU Bloomington. It's all about embracing all kinds of intelligence and ways of studying it, trying to find commonality and connection. They have shown me many perspectives and experts that I greatly respect (and some I disagree with), and might not have found otherwise.
@ngaylinn @neuralreckoning great, I will check them out and the website too, thanks!!
@elduvelle_neuro @neuralreckoning The podcast is excellent, by the way. It's a really good place to start.

@elduvelle_neuro

My personal opinion is that people who claim to see intelligence in animals are flailing badly. Sure, birds can count to seven or so, maybe. But logical reasoning based on concepts? I don't think so. For intelligence in people, some of the philosophers are somewhat interesting. Jerry Fodor and Daniel Dennett and the like. Also stuff like Metaphors we Live by (Lakoff and Johnson), that is, the non-Chomsky linguists who thought about semantics interacting with grammar.

@djl thanks! So do you think non-human animals are incapable of having "concepts"? Or just that their concepts are too "simple" to be used in an intelligent way? What would convince you that an animal is able to reason in a concept-based way?

(Hard questions I know)

@elduvelle_neuro

I'll go with "too simple to be used to get anywhere near human intelligence".

There's some good work to be done on figuring out, for example, what sort of concepts trained dogs (e.g. the ones that can distinguish a lot of words) are using, but I've not seen it done yet. (Maybe I missed it.)

Whatever. Animals (especially cats) are seriously kewl in the extreme. But Lie Algebras and the Standard Model they ain't doing.

@djl OK, fair enough!

In the same way that a human will not be able to perform very well a task that, say, ants might be really good at, we can say that humans will never reach "ant-level" intelligence.. But I'm interested in the general concept of intelligence, not one that is specific to a given species (doing maths is very human-centric). So, if we define intelligence as "logical manipulation of concepts" we should be able to test that in different species.

For example in the spatial cognition field we have the concept of "cognitive map" and different tests of "concept manipulation" like shortcutting tests...

@elduvelle_neuro not what you're asking for, but if you're asking these questions you may be interested in the Adrian Tchaikovsky's Children of Time trilogy (with a 4th book coming next year). The books look at what happens to several species after a series of accidents leave them & humanity in unexpected situations.

They may be something to read while you look for research, at least.