The space of dry table wines has been dominated since at least the 1970s by big red wines and crisp white wines. Big red wines are structured by tannin and alcohol, whereas crisp whites are structured primarily by acid (alcohol might be high in a crisp white, but it tends to lie in the background when the wines are served chilled).

Until the 2000s, in the US at any rate, we tended to regard dry rose as an oddity, but since 2000 the market for dry rose has expanded and we have needed to expand our conceptual scheme for characterizing wine styles. A first approximation would place rose right in between white and red wine (see Fig. 2).

And it is true that the color of rose is in between white and red wine, but a more revealing taxonomy captures the idea that rose is made from dark-skinned grapes, but made using the white wine process that extracts pigments only minimally. This suggests a grid that maps process and grape color to different axes (see Fig. 3).

The grid also makes clear that there is a fourth option, which is to use a red wine process with white grapes. Since 2010 or so, the wine trade has started to become aware of this fourth style, which is traditional in many parts of Eurasia (continuously practiced, e.g., in the republic of Georgia and revived in northeastern Italy and Slovenia). This fourth grid point (see Fig. 4) does not have a consistent naming convention associated with it yet. The Georgians call it amber wine, and that is the name I prefer, but it has also been called orange wine (snappy, but suggests that the wine may be oxidized, may be sweet, or may be made from oranges) and skin fermented white wine (precise but cumbersome). This is a much better way to categorize table wines, and this is the rough and ready classification I hope we start to adopt more widely for the purposes of basic wine education and communication. While it is incomplete, it introduces the two key domains of variation that generate styles: grape choice and winemaking process. At the end of the day, these categories contain the main differentia that set species of wine apart from one another.

To build a slighlty more comprehensive classification, let’s call attention to one aspect of winemaking process: skin contact/maceration, and one aspect of grape choice: must color potential. These are substantial simplifications since they leave out things like barrel usage and maceration temperature and the reasons that the grapes have such and such a color potentioal, which could be early/late picking, environmental/terroir factors, or even blending: a must might have moderate color potential because it is composed of a pale red grape like Pinot Noir, or it may have moderate color potential because it is a blend of grapes with low and high skin pigmentation. Especially in red wines, color potential is a good but imperfect proxy for tannin levels, so we get 1.6 birds with one stone). In general maceration length is a good guide to the extent of extraction. As a result, despite leaving out a lot of details, this classification scheme still captures a lot of impirtant information about the structural qualities of a wine. See Fig. 5.

The advantage of moving from two process options to four is that we can better represent the decision space winemakers face: press immediately, macerate for a while but press before fermentation, press in the middle of fermentation, and press after the conclusion of fermentation. This 4x4 grid better represents the stylistic range that results when applying these four winemaking processes to musts of different color potential. (While color is a decent proxy for tannin intensity, there are outliers: Nebbiolo, for instance, is often more tannic than its color indicates, and Malbec and Syrah are often less tannic than their color suggests.) When we break the grid into regions that correspond to the wines named on the original 2x2 grid, we see that there is an empty space in the upper center of the grid (see Fig. 6).

This L-shaped space is where we find what I think we should call light red wines (see Fig. 7). The Beaujolais region of eastern France gives us the most widely known examples of this style in its lightest-bodied reds, but there are others as well: Pineau d’Anuis from the Loire valley has this character, as do many lighter-weight Grenache-based wines and some Cabernet Franc-based wines; what is revealing is that when we expand the space to four categories on each dimension, we can systematically capture important aspects of diversity that were hidden in the 2x2 grid. Insofar as quality needs to be understood relative to style (something that the beer world has long acknowledged), the 4x4 grid gives us a conceptual scheme that helps us to judge wines by appropriate standards: when we step past the realm of flaws and enter a space of soundly made wines, excellence in a Beaujolais is clearly different from excellence in red Bordeaux. Popularizing a more nuanced conception of wine style is, I believe, going to be an important part of overcoming the idea that wine scores tell consumers the main thing they should know about a wine at the point of sale. This has long been the work of wine buyers, floor sommeliers, and critics (all of whom insist that a score is just part of the picture). There is also work to be done in devising and popularizing simple but compelling conceptual schemes that account for more of the actual variations in wine styles. The conceptual scheme sketched here has the advantage of allowing me to better describe the wines I'm making at Marginalia, but it also demonstrates its merit in giving us a scheme for explaining some key differences between the many styles of rose, red, and white wine.