An experiment without a hypothesis?

This should be a comment, but it will run long. Here's what you note

However I'm surprised that some of the answers have implied that there's something trivial, invalid or unscientific about experimenting without a testable hypothesis: i.e. it's "just playing about" or "just demonstration".

Let me try to explain why there is a problem with a scientific experiment that has no hypothesis.

Assume you are going to do action A to 'see what happens' as a scientific experiment.

There are two possible outcomes:

a) nothing happens
b) X and Y occurred after A

Both outcomes are described incompletely (not scientifically).

In case of a), let us assume that we want to repeat the experiment. To do that, and to arrive at the same result one must know what was observed. However, once you define what you are observing (and how) you have effectively defined a hypothesis.

Similarly, in the case of b), if you do not specify if you were observing for event Z (obviously you were observing if X and Y would occur), the experiment can not be repeated (might yield different results), so the results can not be taken as scientific.

For the results of the experiment to be scientific I think you need repeatability, which must include the definition of the hypothesis and the testing method.

If none is given it is definitively not clear how were you testing (i.e. testing for anything will be different for different people). Therefore I support the opinion that you are not talking about an experiment.

However, as long as you can describe what you were doing (and it was controlled), you can extract the hypothesis that you were effectively testing and turn the experiment into formal one.

If your test was not controlled then it was an accident, as per original Bacon's definition.

EDIT:
You might want to read upon this article, too (do check the references though). Maybe research is better suited in this case.

Maybe you need to expand your understanding of hypothesis a little bit.

If you are doing something just to see what happens, then your hypothesis could be "something detectable will happen" or possibly "the resulting data will demonstrate a meaningful pattern".

One rather dreary quote on a page about scientific experiments said

Any laboratory procedure you follow without a hypothesis is really not an experiment. It is just an exercise or demonstration of what is already known.

You might call an experiment without a hypothesis a scientific demonstration. The Wikipedia article explains:

A scientific demonstration is a scientific experiment carried out for the purposes of demonstrating scientific principles, rather than for hypothesis testing or knowledge gathering (although they may originally have been carried out for these purposes).

Some famous demonstrations that are mentioned are:

• Al-Biruni's reaction time
• Alhazen's camera obscura, lamp experiment and magnifying lens
• Al-Jazari's crankshaft, elephant clock and programmable robots
• Avenzoar's parasites
• Detonating a cloud of flour
• Foucault's pendulum

You could say one was just fooling around in the lab. For a much less serious phrase, I would look at the term (informal and may be construed as offensive) futz:

: fool around 1 —often used with around. futz around without producing any worthwhile music — John Koegel

If you wanted to describe just walking into a lab and seeing what happened, you could say:

On my day off I spent some time in the lab. I didn't have any projects, so I was just futzing around with pulleys to see what would happen.

Edit: One more serious option, as brought up by both @Richard and @Jay: scientific exploration. This brings to mind someone in a lab trying various things out in order to "see what happens". If your purpose is to be more formal, I would use this term. If you are, however, joking about playing around in a lab, I would still use "futzing".

I know your question is about scientific hypotheses in general. But in the domain of software testing, a test case proceeds as a scientific experiment. You have a hypothesis, i.e., that under certain circumstances the software will behave in the specified manner, and you exercise the system and observe the output in order to prove or disprove that hypothesis (making the test "pass" or "fail", respectively).

But sometimes, testers execute test cases without any particular hypothesis, either for learning purposes or to refine the quality of the tests. This is called exploratory testing.