智能不等于意识：一篇关于智能本质的猜想

I have a conjecture which I’m almost certain is true. Intelligence — as defined as ability to solve open-ended, novel problems — Yes it’s my em-dash if at all necessary, is not strictly the same as sentience, which — without defining it — consists of more than intelligence. Even if intelligence is a necessary ingredient for sentience (which I don’t think it is), sentience is a strict super set of it.

Just like how it’s possible to say a vessel is larger than another vessel contained within it, without having to identify what other things are in the larger vessel, we can say humans (who are definitionally sentient) and other animals contain intelligence, but they are not just intelligent.

Though I’m not certain as to what constitute sentience (if it constitutes anything at all), if I had to hazard a guess, Agency would be one of my picks. Even inanimate things can be imbued with agency, but only sentient things can provide the first touch to infuse it onto them.

Personally, I think abstract intelligence consists of a very small set of primitive capabilities. Similar to computation, intelligence is probably a universal characteristic that can show up in unexpected places that have those primitive capabilities. Today we aren’t surprised when we realize sometimes card games and tiles unintentionally have the same expressive capability as any modern computer. Similarly, we shouldn’t be surprised when we see intelligence show up in systems that have the set of primitive capabilities required for it. It’s possible that either these capabilities or their configuration might be rare, which might explain their relative scarcity. Nonetheless, just like algorithms have some core primitives that allows them to describe computations, intelligence might have core primitives. The mechanism to create an algorithm with primitives to describe a computation itself might have other core primitives. Seen this way, intelligence might just be a higher abstraction of computation.

Moreover I think intelligence is just a very small part of what is required to solve a problem. The major part of solving any problem is the computation required to execute on the spark of intelligence. Apart from the strictly mechanical following of an algorithm to get a result, here I use the term the computation more broadly to mean the ability to evaluate and pursue novel paths that might open up during exploration based on experience, adjusting techniques to the scenario and so on, essentially anything that is a transformation and fusion — however novel — of existing ideas. From this perspective, a lot of ideas might be seen as a remix. My crass categorization is not in any way meant to diminish the value of computation as I have defined it. On the contrary, this work of navigating the idea maze is the primary activity that provides us with the actual fruits of intelligence be it theorems or blueprints on how to build bridges.

I do not offer much in the way of concrete evidence regarding this view. If pressed I might point to the somewhat universal nature of intelligence in all animals and plants, and the fact that hints of intelligence (at least in the sense I’ve defined above) seems to have popped out of what seems like arbitrary operations in machines. I acknowledge that even if it’s considered evidence at all, it’s very scant. Hopefully, they at least offer some glimmer of universality of intelligence, even if it just turns out to be a mirage.

Weakly, I also think it’s possible that intelligence is a binary characteristic i.e, it’s either present or not. I feel this partly because of my aforementioned belief that intelligence is comprised of a bunch of primitives capabilities which I think are either present or not. Qualitatively, it doesn’t feel like even the most intelligent people are that much different than the average and the variance can probably be explained by their ability to persist and search through a problem space much more broadly Long back, there was this tweet, where the author was asking what evidence is there that humans are not at the peak of intelligence curve, although I think it’s very pretentious to assume we are the pinnacle of intelligence, the idea stuck with me. I’m reminded of it now in the current context, because it can be reinterpreted to mean that all intelligent entities are essentially equal in terms of intelligence and the divergence arises only due to the computation part of intelligence. Unfortunately I can’t find the link for it, despite trying to find it. . It can probably also be explained by having a larger and more diversely cultivated bag of tricks.

Language is probably a very costly wrapper around the core of intelligence. The advantage of language however is that you can horizontally scale problem solving. I don’t think language enhances the core intelligence in any way, but even on an individual level, it helps concretize the artifacts of intelligence to be recorded and referenced with words and symbols. More commonly, it opens up the problem to be shared between multiple people, so that the computation and the incidental problems that actual require intelligence can be done by multiple people in parallel to solve the task.

Those are a lot of unknowns and conjectures, my hope is that a new field emerges that would study intelligence with the same rigour and formalism as computation and mathematics, and that this field would be studied abstractly without being tied to an individual organism be it a man or an animal.

All of this is to say, I don’t think intelligence or language are the essence of what makes us sentient. Intelligence, is likely a universal characteristic, that humans seem to have latched on to as fundamental to their identity as Homo sapiens and seeing this show up elsewhere is obliterating that sense of identity and causing a lot of agony. This isn’t helped by the fact that extremely motivated people are making bad faith arguments and pushing complex babbling machines as sentient which need to be schooled in morality by employing religious leaders.

You might not agree with anything I said above about intelligence and sentience, I acknowledge there are other various definitions for intelligence and sentience. But a lack of clear definition is not a barrier to delineate between two distinct concepts. Given that things in nature don’t come tagged with abstract concepts, these delineation of abstract concepts are necessarily artificial. But we don’t have to be too post-modern about it to not know the difference between a submarine and a fish.