If human language is really superior to animal communication, then why can't we all easily learn to speak cat? Or dog? Or dolphin?

And I mean that as a serious question. If humans have animal communication plus added complexity (recursion or whatever you want to claim), then doesn't that mean we should be able to also still speak cat, dog and dolphin?

A system that learns one thing doesn't necessarily learn every simplification of that thing. This is because a more complex hypothesis space is going to be larger, and so learning a simple system with that complex hypothesis space requires two kinds of evidence: 1) evidence that the system is simple, and 2) evidence about the behavior of that system. A learner that does not consider the large hypothesis space needs only the second kind of evidence. Indeed, a complex learner might assume that the system is complex, in which case no amount of the first kind of evidence will be enough.

For a concrete example, we can view context free grammars as a generalization of linear structures by saying that a linear structure is a tree where every local subtree branches to the right:



A learner that assumes a linear structure on data that has only linear structure will discover the linear regularities very quickly. A learner that assumes that local subtrees may branch in either direction, however, will need both evidence for the linear regularities and evidence that there is nothing except rightward branching structure. Moreover, if the learner refuses to consider the possibility that all trees branch to the right, then the learner will never learn the correct linear structure.

All species have special things that other species don't do - that is why they have their own niche and adaptive pressures resulted in them speciating in the first place. Humans, like other animals, have unique traits.  I believe one of these things is language, but there are a number of other cognitive skills humans have that are not found in other animals. Language isn't monolithic, and is made up of a huge amount of interacting skills - and these skills can be found in other species to varying degrees sometimes.

But humans are the only ones that intentionally communicate with a learned system of symbols that have been agreed upon by a community. And humans are the only animals that live in this niche, with this type of social structure and life strategies. It's not superior, because 'superior' is a subjective designation and is in this context meaningless, like discussing which communication system is more beautiful.

For my research I'm looking at the relationship between theory of mind and language.  Theory of mind is also a trait that is rare and possibly unique to the human species.  Both of these traits scaffold the development of each other and are required for the other to work properly. There are a number of developments that are required before the development of theory of mind or language - and it's the lack of these, as they haven't been evolutionarily adaptive and have not developed in other species, that other species can't be taught to speak, and we can't be taught their communication system. So it's a much more complicated story than that language is just 'communication plus complexity', or our system with a little less complexity and you get 'cat'.

I don't find the term 'language faculty' helpful at all - I suggest tossing it away and things might be a bit easier to talk about!

One thing that I always come to think of in arguments like this is that other animals don't require language to the degree that humans do. Whether there is a biological/neurological aspect of human language or whether other animals lack this, is irrelevant — other animals do not need language like we do.

If any other species developed the same need for communication, what is there to say that they wouldn't develop something similar to human language, or that it couldn't be analysed similar to human language?

These are quite big assumptions being made, all hinging upon some axiom that human language is "special".

If you compare pit hole eyes to vertebrate eyes, one could say that the pit hole eyes are more primitive, but if you compare fish eyes to human eyes, which one is more primitive?

And do all animals need vertebrate eyes? Why would vertebrate eyes be special?

While I agree that there may be issues of whether species need language or not, we can't assume that all have the potential for it.

Of course, not!

It may develop specifically when a species does need it, for one thing.

Which is why it seems so strange, why judge animal communication based on our form of language, based on our evolution, when no other animal seems to have developed any similar form of language?

That in itself seems assumptive.

jkpate, that's an interesting point. However, wouldn't that mean that we'd at least be able to understand cat? Maybe we would inherently speak it in a way that is too complicated for other animals to understand, but a simpler system would be within what we would understand, as part of understanding the more complex system. I think.

I don't think so. If a learner assumes that there is hierarchical structure when there is not hierarchical structure, then the learner could produce systematically wrong parses. If accurate parses are necessary for understanding, then such a learner would not understand the system. Whether this happens would depend on exactly how the learner is set up (whether it has a strong bias for having branches in both directions, for example) and what the data look like (in terms of amount and the likelihood of spurious hierarchical regularities).

More concretely, do any animal communication systems engage in metapragmatic functions? I'm not aware of any that do, and it's hard to imagine a more defining feature of human language use (cf. threads like this one).

Humans are indeed above the beasts, simply because we proclaim that we are. Interesting thought.

I'm speaking a little more broadly, but I also don't know of any species other than humans that discusses such things. I'm skeptical about that implying any sort of "superiority" but you certainly do seem to be right that metapragmatics (along with other kinds of self-analysis) seem to be lacking in other animals, at least as far as we know.