I've written a few wine reviews attributing flinty or chalky "minerals" in wines to the soil in which they are grown, but the scientist in me has always puzzled over how the grapes do that.
The answer is, quite simply, that they don't. Weather, soil chemistry, and soil structure, of course do play a role in gene expression and thus in the sensory qualities of wine, but the idea that the vines slurp up aroma-of-rock and excrete it in grapes is pure fancy.
Randall Graham a winemaker at Bonny Doon, reportedly experimented with adding rock to wine the way some add oak chips. Results were somewhat amusing, with the rocks buffering the pH and drastically altering both the taste and mouthfeel.
Harold McGee and Daniel Patterson, writing in the New York Times magazine attribute terroir more to the local winemaking traditions than to the soil. I find the argument rather convincing, that it is the winemaker and not the dirt that is steely or chalky, but I'm not as completely sold on it as Patterson and McGee. Yes, the grapes are not expressing the taste of the dirt, but topography, rainfall, soil structure, and composition play more than minimal roles in determining character. Some of the lesser-known "cult" wineries and upstarts will bring out pours of unblended wine from different vineyards to more enthusiastic tasters.
I recall a very proud owner-operator in Livermore giving me a sip of this and a sip of that (supposedly vinified the same way), then pouring three Zinfandels blended largely from the vineyards. All Zin, and all quite different. Granted, they could be from different clones; I didn't think to ask.
Terroir is still a mystery. What is genetic, what is environmental, and what is winemaker has yet to be pinned down. Even the question of what we mean by "minerals" in the wine--it isn't a higher salt concentration!--remains unanswered. Some might like the mystery to remain, but as for me, these investigations enhance my enjoyment.