What is the proper definition of "sound"?

Framing this from white's perspective, is an opening considered sound only so long as it retains the advantage (or, at worst, concedes equality)? Or would an opening still be considered sound if it gives a slight advantage to black? Or what exactly is the demarcation between sound and unsound: +/=, =, =/+, -/+ or even -+.

I think an opening is considered sound as long as the opponent doesn't have a forseeable advantage by its end. The half point advantage that white supposedly has (although I disagree with this theory) isn't really a factor.

Yea, some "gambits" are considered unsound because they have been completely refuted.  As long as the plus score is pretty good, something is still considered sound.

That's an opine, and the way I understand it, in a nutshell.  :)

I believe that an unsound opening is one that gives at least a +/- (or the equivalent) to the opponent, such as the Damiano Defense.

It's sound if it doesn't have an ourighr refutation, imo.

1 b4 is sound, 1 e4 e5 2Nf3 f6 is not

I was going to give the same answer as some above: If it's refuted it's not sound.

Then I figured I'd have to define refuted, and not made any progress at all.

I can live with refuted as a pretty self-evident term.

I do wish the masters would define their terms better sometimes though. I would say that an opening is unsound if it can be clearly shown to lead to a +/- or -/+ in all variations. Anything less than that I would label dubious or questionable maybe, but not "unsound".

Well, apparently it's not when a tree falls and no-one is around to hear it.

I suppose so - if you are correct. I'm not up on my Halloween Gambit or Smith-Morra.

Defining unsound is a fun exercise.

In my own limited experience, I usually reject as unsound any opening in which my opponent can forcibly reach an evaluation of -/+ against me. Admittedly, that's a pretty liberal definition. One that no GM would ever use.

But then that raises the appropriate question: are some openings unsound for super-strong players but less so for novices or club players?

One side of that argument suggests that you should never waste time learning openings which don't confer White's edge to him or which don't give Black the best chance at equalizing or finding considerable counter-play.

The other side suggests that there are many reasons for playing openings which seem to be refuted (or at least highly frowned on) at the GM level. These might include openings which allow for very tactically oriented play from the earliest moments, or which provide initiative that one must not learn to squander, or which provide the learning ground for holding an endgame a pawn down.

The two sides will never come to agreement on this one, I don't believe. But many openings thought to be long refuted (at least the winning chances refuted) have come back and used by the very best as very powerful weapons - at least until the new refutation arrived. (Two examples come to mind: a) Kasparov's Evan's Gambit win at Riga against Anand in 1995, and b) Movensian's Max Lange win at the 2009 Corus versus Michael Adams.)

Perhaps a token offering of peace between the two sides of this argument can be had by suggesting that juniors, club players, and learning players in general should certainly consider openings like the Bishop's Opening and the King's Gambit (neither of which is unsound but neither of which make up the chief weapons in a super GM's opening repertoire). Indeed, the Polgar sisters grew up on these openings, cutting their eye teeth on them as well as the figurative throats of their opponents.

And a final note: you'll never convince afficianadoes of the Blackmar-Diemer Gambit to back away from the use of their unsound gambit, not as long as they keep winning games with it (though again, there are precious few GM games with such lines).

So, it's evident to me, at least, that the notion of unsound is not so easy to define after all. I'm still having trouble with it, for sure.

Dictionary: sound¹   (sound)

n.

1. 1. Vibrations transmitted through an elastic solid or a liquid or gas, with frequencies in the approximate range of 20 to 20,000 hertz, capable of being detected by human organs of hearing.
   2. Transmitted vibrations of any frequency.
   3. The sensation stimulated in the organs of hearing by such vibrations in the air or other medium.
   4. Such sensations considered as a group.
2. A distinctive noise: a hollow sound.
3. The distance over which something can be heard: within sound of my voice.
4. Linguistics.
   1. An articulation made by the vocal apparatus: a vowel sound.
   2. The distinctive character of such an articulation: The words bear and bare have the same sound.
5. A mental impression; an implication: didn't like the sound of the invitation.
6. Auditory material that is recorded, as for a movie.
7. Meaningless noise.
8. Music. A distinctive style, as of an orchestra or a singer.
9. Archaic. Rumor; report.

An unsound opening is one that leads to a clear advantage for the other side with minimal amount of theoretical knowledge by the opponent.  For example, if I find a 40 move refute to an opening, I wouldn't mind playing it because I know the other side won't find the refute.

Sound

Correct. A sound sacrifice has sufficient compensation, a sound opening or variation has no known refutation, and a sound composition has no cooks.

This is from the chess glossary on wikipedia.

I remember once asking a very strong IM if the latvian gambit was a "bad" opening. His answer was that there are no bad openings, only bad chess players. 

Something is sound when it can't be refuted outright. It's unsound when it probably can be -- so either a refutation is known, or its correctness hangs on such tiny threads that nobody would be surprised if a final refutation was found.

For a white opening, if it's known to lead to a small advantage for black with best play, it's refuted.

 You stole my idea!

Thanks for the extensive write-up on the subject DrawMaster. It's uncommon for most to put that much effort into musing over and writing on a subject.

I pretty much agree with all you said.

I've never before heard the stipulation about minimal theoretical knowledge.

I tend to agree with your IM's comment. Gambits especially, even dubious gambits, seem often to be fine practical weapons against most competition.