Folks frequently refer to an application’s features as intuitive or unintuitive. But I’ve heard the idea from different quarters that in computer technology there really is no such thing as an intuitive design because the interfaces are so far removed from the physical & ecological reality on which human intuition is based and developed over millennia. What do you think?

Intuitive interfaces can only be found through following the rule of least surprise, they do not exist through any intrinsic property. For example, I find programs modeled after the text editor vim to be very fast and intuitive, however, a user unacquainted with vim would find them incredibly dense and unintuitive. So, although it is the job of the designer to remove anything which would actively misdirect the user (in vim, pressing “d” deletes text), as long as it conforms to expectations of behavior, it can either be seen as intuitive or become that way through practice.
I’m on board with Chas — I think an interface is intuitive if you know how to use it without having to receive instructions.
There’s a chicken-or-egg thing here: maybe a checkbox on a form wasn’t intuitive to anyone the first time it appeared but it’s so ubiquitous that it becomes intuitive over time. Fair enough. That’s how we learn to operate in the real world, as my 2-year-old constantly reminds me (”no, don’t touch the fire ants…”)
I think that we’re referring to a lot of problems with such a simple question!
The first point, easy to overlook is the man-machine interface: the keyboard and the mouse… or the touch screen.
The physical interface is the first stop, that could be ignored just when we agree that we are talking just with “experienced” people (that’s a problem I found when designing interfaces for banks and governments, since they need to talk with EVERYBODY).
Even if we consider the people experienced in this first step, we have to consider that different OSes have some minor differences. To cite one: the position of OK and Cancel buttons (right – left).
When designing web apps I try to abstract that out trying to avoid metaphors that triggers those behaviors (for example, preferring the step-by-step wizard metaphore, with just forward and back).
Ok, too much for just an intro: after those points the first problem is about mental models. We need to find all the common mental models and habits that are used by our “target” people. A part of this task is addressed by some kind of ethnographic research, and the other (more technological) is about finding the way to attach to what I called “Circadian Activity Flow”, the sequence of actions of a person during the lapse of a day.