Coloring white wine red can shift even the words used to describe its aroma toward those associated with red wine. In a 2001 study, 54 enology students tended to apply aroma descriptors they had previously used for red wine to a white wine that had been colored red.
The paper was not solely an experiment with colored wine. It began with a vocabulary analysis of a large collection of professional wine reviews. The study explored how visible color connects with the process of putting aromas into words.
Words describing wine aromas separated by color
The researchers analyzed four collections of professional tasting comments: 3,000 French-language magazine reviews, 3,000 reviews selected from a guidebook, 7,000 English-language wine reviews, and 2,000 entries from a French winemaker’s personal records.
The analysis statistically grouped words that tended to occur together in the same texts. Wine color, region, vintage, and score were then related to those groups. The procedure did not begin by assigning words to red and white categories.
The number and content of vocabulary groups differed between collections, but the first major division consistently separated red wine from white wine. An analysis combining the three French-language collections also produced two word groups corresponding to red and white.
The authors then considered the colors of objects named in aroma descriptions. White wines were often described with objects suggesting light or yellow colors, such as lemon, honey, butter, white peach, and pear. Red wines were often described with red or dark objects, such as cherry, raspberry, blackcurrant, and chocolate. This was reported as a tendency shared across all four collections, not a rule applying to every word.
Checking the dye’s effect before turning white wine red
A connection between vocabulary and color alone cannot show whether red and white wines genuinely differ in aroma compounds or whether visible color is responsible. The researchers therefore conducted an experiment changing only the color of the same white wine.
They used 1996 Bordeaux wines: a white containing Sémillon and Sauvignon grapes, and a red containing Cabernet Sauvignon and Merlot. Part of the white wine was colored red with 2 g/L of purified grape-derived anthocyanins. The uncolored white wine underwent the same stirring and oxygen exposure.
In a preliminary control test, 50 people compared the colored and original white wines in opaque glasses under red lighting. Each completed six trials selecting the odd sample from three, and the authors reported no significant discrimination. This check investigated whether the dye substantially changed the aroma; it did not prove that the two samples were completely identical in every sensory property.
Reusing their own aroma vocabulary a week later
The main experiment involved 54 students from the University of Bordeaux’s Faculty of Enology, with equal numbers of men and women. It was not an experiment composed exclusively of renowned sommeliers. Participants compared two wines in clear glasses under white lighting, allowing them to see their colors.
The first session used the original white and red wines. Participants could use supplied descriptors or add their own, and indicated which wine expressed each aroma characteristic more strongly.
One week later, they were presented with the original white wine and the same white wine colored red. Each participant used the list of descriptors they had personally selected in the first session. The words were reordered alphabetically, with no indication of which wine they had previously described. Participants again indicated which of the two samples expressed each characteristic more strongly.
A distinctive feature of the design was that it tracked each person’s use of words across the two comparisons without imposing a shared “correct” vocabulary. It is equally important that the second session was not entirely free description: participants reused their earlier terms.
Red-wine descriptors shifted to the red-colored white wine
In the first session, descriptors such as honey, citrus, and pear were used for the white wine, and blackcurrant, raspberry, and cherry for the red. In the second session, participants generally applied the terms previously used for red wine to the white wine colored red. Conversely, descriptors previously used for the original white were less likely to be applied to the colored sample.
In other words, even when comparing the same white wine, aroma descriptions separated along the color difference. The connection between wine color and aroma vocabulary found in professional comments also appeared when color was manipulated experimentally.
However, the task did not directly ask participants to identify whether a wine was red or white. It asked which descriptor fitted which sample better. Describing red-colored white wine with red-wine vocabulary should be distinguished from claiming that every participant misidentified the beverage in the same way.
The authors focused on the stage of putting aromas into words
The authors interpreted the color-related changes in aroma descriptions as a perceptual illusion involving vision and olfaction. In particular, they discussed the possibility that visual information enters at the “verbalization” stage, when an odor is translated into an object’s name.
For example, seeing red might make red-fruit candidates easier to bring to mind, prompting an ambiguous aroma to be described in those terms. The paper discusses visual mental imagery and memory representations of odors and words, but the experiment itself did not directly measure brain activity or mental imagery.
The central finding supported by the study is a change in descriptors accompanying a color manipulation. The design did not fully distinguish whether that change occurred in immediate sensation, odor identification, retrieval of words from memory, or response selection. The authors’ broader ideas about olfactory function are best treated as hypotheses arising from this limited task.
Conclusion
Professional wine reviews showed a major division in aroma vocabulary between red and white wines, with the colors of named objects also tending to correspond to wine color. In the experiment with 54 students, coloring white wine red made descriptors previously used for red wine more likely to be applied.
When putting aromas into words, we use not only information from the nose but also cues from visible color. Both the analysis of a large body of text and the color-manipulation experiment demonstrated this connection.
The study was supported by France’s Centre National de la Recherche Scientifique (CNRS), Institut National de la Recherche Agronomique (INRA), and Bordeaux 2 University. No separate conflict-of-interest statement was found in the original article.
Reference
Morrot G, Brochet F, Dubourdieu D. The Color of Odors. Brain and Language. 2001;79:309–320. doi:10.1006/brln.2001.2493
