The first sentence of the abstract is "When dogs interact with humans, they often show appropriate reactions to human intentional action. But it is unclear from these everyday observations whether the dogs simply respond to the action outcomes or whether they are able to discriminate between different categories of actions"

They don't assume they are dumb, they just don't assume anything, as they should.

What other assumption can you start with?

Assuming the natural world is dumb is simple, consistent, and easy to explain.

If you start with the assumption the natural world is “smart”, what does that even mean? Does it mean that we should assume dog can do calculus, but choose not to? Or that dogs are capable of complex communication, but also choose not to?

Saying the natural world is “smart” is a pretty meaningless statement. How smart is “smart”, everyone will have a different opinion, so you can’t use it as the basis for discovery or discussion.

However everyone understands what the natural world being dumb means. From there you can start the journey of proving step-by-step where the smarts are.

I don’t think that the assumption is that the natural world is dumb. The scientific discovery process aims to highlight new insights to help us make a better and useful description of the world. Implying that the starting point is "dumb" is incorrect. The starting point is a blank page, that will be filled by the research, whatever they find out.

I understand what your trying to say. But I don’t agree with words you’ve used.

We don’t assume nothing, that’s clearly not true. A basic assumption we make is that the natural world is capable of perceiving us and responding to our actions.

The question we’re normally testing is how complex the process between perception and response is. As general rule we assume that process is extremely simple (“dumb”), and work to understand how complex (“smart”) it actually is.

For example, in this paper the scientists didn’t start by proving the dogs were capable of observing the humans and responding to their actions. That was a given. What was tested was how complex the dogs perception and decision making process was.

Using words like “dumb” and “smart” are extremely crude. But it’s the words used by GP and OP.

It's the opposite : you assume that the natural world is smart, but they don't assume anything, they just test the hypothesis that dogs are smart, and in this way validate part of your assumption.

I think it's a better way to deal with things, as your assumption might not hold scrutiny very well : what is natural world ? Are worms part of it ? What about rocks ? Can we deduce from there that worms and rocks are able to distinguish intentional from non intentional actions ?

If you have ever written a game KI or something alike:

the most simple algorithms (few lines of code) can perceive the observer into thinking there is advanced KI at work.

Same might be the case with natural intelligence.

What does KI mean?

It means AI in German (Künstliche Intelligenz).

I never made that art / artificial connection before, but in retrospect it's obvious.

Finding wordroots is fun and sometimes illuminating.

Really? Because it seems to me that they suspected the opposite - that animals are intelligent - and they designed an experiment to attempt to prove it.