Postmodern Winemaking


Astringency And Harmony In Tannins

September 2010
Just as the heart of rock and roll is the beat, the foundation of the soul of wine is texture. In this column, we have often discussed the connection between fine colloidal structure and the aromatic integration that leads to soulfulness. Beyond this physical feature, great wines are tuned in to harmony, a human preference that we hold in common. For the postmodern winemaker to work in this medium requires both technical expertise and artistic sensibility.

During the past two decades, researchers have developed a language for tannins and a roadmap for their evolution. Much technique lost in the modern era has been recaptured concerning the role of oxygen, lees and oak; additional tools such as membrane technologies have been added to the winemaker’s kit. In my previous eight columns, I have laid out the resources now at the disposal of the postmodern winemaker and outlined the consequences of various approaches.

No general dogma exists guiding the degree of refinement, edge or rusticity that a winemaker may deem appropriate. For style direction, the practitioner is well advised to consider traditions for which the local appellation has or could become known, with an eye to minimizing consumer confusion by delivering wines that support a consistent local expectation. Depending on the region’s historical context, the winemaker’s experimental latitude will vary. A winegrower in, say, the Suisun Valley of California may be advised to swing out more creatively than might be stylistically appropriate in the more established neighboring Napa Valley region, and might even be contrary to local ordinance in the Médoc.

Nature in the rough
Why does astringency exist? Primarily, it is a way for plants to persuade us not to eat their children. Actually, they would be delighted for us (and birds) to swallow their children whole, but not to chew them up on the way down, until they are mature and ready for the voyage through our entrails and into fertile soil.

Astringency is a tactile sensation, not a taste. As hunter-gatherers, our forbears needed sensory cues to phenolic content -- astringency and its taste analog, bitterness -- to distinguish food from poison, to determine ripeness and so forth. Sensory stimuli from all five senses are directed to the thalamus in the midbrain, where snap decisions tell us whether to swallow or spit. The midbrain is like a security kiosk at the gateway to our cognition, and our job as winemakers is largely to get the thalamus to stamp our passport and send us up to the frontal lobes instead of down to the fight-or-flight centers of the reptilian limbic system.

Tannin can confer richness and satisfaction, and its integrative properties can enhance soulfulness, but excessive astringency is repulsive and hides the wine’s charms. Threading this needle requires a great depth of cellar skills. Since winemaking is a performance art, we must also know our audience and the extent to which it differs from us.

Groping the beast
Richard Gawel, of the Australian Wine Research Institute’s Tannin Project, published in 1998 an excellent review of red wine astringency research.1 Its physical basis is the natural affinity of salivary proteins for wine phenols, a combination of hydrogen binding and hydrophobic interactions. These are weaker than true covalent bonds, so the longer a tannin polymer becomes, the stronger the cooperative binding to protein.

Doug Adams’ astringency assay2 shows that protein binding begins with polymers about four units long.

Postmodern winemaking makes use of these principles to shape polyphenols that are as small as possible, thus forming into small colloids that maximize surface area and promote hydrophilic/hydrophobic interaction. Incorporation of anthocyanins is thought to terminate polymerization: The greater the color, the finer the tannin. Interaction with salivary proteins can be blocked by the incorporation of lees peptides and other sulfur-containing compounds as side chains on the polyphenols, thus reducing harshness. With care, oxidative linkages can bind sulfides to tannins, converting stinkiness into softness.

Duel of the titans
Astringency’s physical causes recently have received much attention from various camps with apparently contrary findings. At the University of California, Davis, Ann Noble’s long and impressive research3 on this topic did not distinguish different kinds of astringency but concentrated instead on exploring differences among tasters in the duration and intensity of astringency that she and Uli Fischer4 linked to individual variations in salivary rate. Likewise, wine rockstar Tim Hanni5 used mapping of tastebud density to develop a theory predicting wine style preferences based on sensitivity.

Richard Gawel’s team, by contrast, produced a Mouthfeel Wheel6 that distinguishes seven primary and 33 secondary terms for astringency. A greatly simplified version of tannin terminology I use in my daily work with wine tannins uses just seven terms and has proven valuable in predicting tannin behavior and progression over time as, hypothetically, phenolics rearrange themselves into polymers and colloids. In addition, my work with alcohol sweetspots, musical pairings and cork taint due to TCA has shown that there is another source of astringency that is not related to composition.

The findings of Noble and Hanni are important in liberating personal preference from the province of pundits. Consumers are entitled to their own likes and dislikes, and they ought to place no more faith in wine reviews than they do in critiques of films and restaurants.

However, taken to the extreme, this disconnectedness leads to the death of art. How can I make delicious wine if my sensory apparatus bears no relationship to that of my customers? If we are all unknowable, then music and even good cooking are pointless endeavors. But experience clearly tells us otherwise.

Fortunately, it is possible through an understanding of the stages of perception to reconcile these two perspectives. To ally these views, we need to distinguish what is actually there, what our senses tell us, what our brain makes of this information and whether we like it. Let’s call these four stages I. Phenomenon, II. Sensation, III. Integration and IV. Preference.

Our two camps form a nice complementarity in this hierarchy. Research distinguishing how people are different concentrates on phases II and IV: differences in sensory hardware and differences in preference. Those who concentrate on strongly shared aspects are more interested in Phases I and III: correlating differences that exist in nature with linguistic descriptions of the sensory impressions the integrative process compiles.

Individual differences in sensory acuity do not divide us as much as they might. We all have different hearing sensitivities, but unless we are profoundly deaf, this doesn’t alter our ability to draw emotional content from music. We all find a major chord cheerful and a minor chord melancholy.

The reason differing sensitivities don’t necessarily preclude finely synchronized aesthetic experiences is that we all share sophisticated compensating software. Once our five senses have collected signals, a huge amount of processing takes place to correct for differences in environment and acuity. Our brains are highly skilled at extracting salient details held in common and compiling an integrated package through leaps and guesses. That’s how we recognize a face years later and from a different angle. We are masters at filling in the gaps. This software is so powerful we don’t even notice we’re doing it.

Humans are born preloaded with software to fill in gaps, something deep in the DNA and strongly shared. This is the source of our sense of harmony, and it explains why fine nuances of discord and harmony, to which we relate without being told, show up in wines as they do in music. Wine’s ability to communicate these subtle messages is its primary appeal, and accessing this mysterious language is the foundation of postmodern winemaking.

Talking tannin
I’ll summarize the postmodern picture of what’s going on. Young red wines generally contain monomeric anthocyanins and unpolymerized tannins such as catechin and epicatechin from skins and seeds. These are not soluble but gather into copigmentation colloids,7 tiny beads that are perceived as a graininess on the tip of the tongue -- tanin vert or “green” tannin. There is at this stage little astringency elsewhere on the palate, even for big wines.

Since I’d been raised with the theory that only polymers are astringent, I asked Montpelier’s legendary phenolics expert, INRA’s Michel Moutounet, why red wines, being primarily monomer, are astringent and then smooth out as they polymerize. He offered the theory that cooperative binding of salivary proteins occurs along the surface of the copigmentation colloids into which the monomers are arranged.

Left undisturbed, these monomers undergo non-oxidative polymerization into regular chains (for you phenolic geeks, linking the 4 and 8 positions on the flavan-3-ol). The resulting polymers are compact and have relatively little mouthfeel and protein interaction. In these conditions, however, anthocyanins are not readily preserved, so the chains continue to elongate, becoming increasingly insoluble and drawn to salivary protein, so that the textural impression after a few years migrates from the tip of the tongue to a general dirty graininess throughout the tongue and cheeks that we call tanin sec, or “dry” tannin, the only type of tannin to occur under the tongue in the back of the mouth.

A small amount of oxygen early in the wine’s life will send these phenols on a different course. Oxidative polymerization, which utilizes the vicinal diphenol cascade I described in February,8 efficiently captures and stabilizes anthocyanins and creates randomized open, freely rotating polymers that interact strongly with saliva, creating an impression of increased volume and aggression. In a few days of micro-oxygenation, green tannin is transformed entirely to tanin dur (“hard” tannin that sits entirely on the top of the tongue and moves back to create an angularity in the finish). The impression is not at all grainy but sheet-like and grippy, like peanut butter, causing the tongue to adhere to the top palate. The oxygen pickup from simply barreling down young red wine moves it in this direction.

Because the oxidative cascade initially increases oxidative reactivity, the formation of hard tannin is accompanied by an increase in reductive strength, and often it will induce formation of sulfides and decreased aromatics. Vintner Randall Grahm of Bonny Doon refers to this period as “the valley of the shadow of death,” a frightening time for the novice practitioner.

After an extended period, however, the grip begins to abate and the wine begins to open. We call this the “firm” stage. Further oxygenation will melt the tannins, beginning from the back palate, where angularity is replaced by velvety softness, while the front of the mouth remains firm, a stage we call “round.” If we are employing MOx, this is often a good place to stop, retaining enough reductive strength for subsequent barrel and bottle aging, while ensuring both stability and longevity. For wines intended for early consumption, oxygen can be continued until the tannin is completely “melted.” Melted tannin is quite stable in the bottle, but it has little anti-oxidative power and can fall apart if excessively exposed, leading to dryness and precipitation.

It’s worth noting that these changes in texture occur without any alteration in the total phenol content of the wine. It’s the way the tannins are assembled that determines their sensory properties (as well as affecting the expression of other aromatics such as Brett and bell pepper). Choices concerning the degree and type of astringency are the responsibility of the winemaker, without which expression of natural terroir is often defeated.

Our seventh term, which refers to oak tannin, is borrowed from Gawel’s nomenclature. Located on the front part of the tongue just behind the tip, this type of tannin is very drying, but that term is already in use for an entirely different type of tannin. Oak is also associated with eugenol, or oil of clove, a topical anesthetic that numbs this area, so we use the term “parching/numbing.” Oxygen can be used to soften this tannin, but take care to start treatment before the wine’s grape tannin profile is resolved, or else dryness may result. It’s easy to slip up in the timing of introduction of oak tannin; wine does not penetrate deep into oak barrels to extract pithy oak tannin until the later stages of oa k cellaring.

The mystery deepens
“Finally, a full understanding of astringency cannot be achieved without an appreciation that it is a perceptual phenomenon. The role of the perceiver cannot be discounted.” So Gawel ends his review.

We have already discussed the role that differences in taste sensitivity, salivary rates and past experience may play. But beyond this lie some very peculiar traits that give us a window to wine’s true nature and our own nature as humans. It is odd, for example, that 4 parts per trillion of TCA can cause a wine to be astringent. The human nose is incredibly sensitive to this moldy aroma, which ruins the fruitiness of wine. Yet at this concentration, it cannot really be binding to salivary protein to a perceptible extent.

TCA’s astringency demonstrates the role that fruit aromas play in creating the illusion of harmony. Consider for a moment the difference between music and noise. At the beginning of a symphony, the instruments quietly tune up -- a most obnoxious sound. Then they strike up with much greater volume and we are carried away by music, for which we have a very high decibel tolerance as long as we are emotionally taken up -- one false note and we wince.

There is a sweetness to music that is bestowed by the midbrain -- its hall-pass to cognition’s upper floors: contemplation, emotion, sometimes ecstasy. Without this endorsement, it’s just a lot of clatter. Good wine is granted similar privileges, and these can be revoked if a disharmony such as corkiness is introduced.

What forced me finally to grasp that wine is liquid music was the phenomenon of sweet spots9. When our company started offering alcohol-adjustment services, wineries came to us because, at high alcohols, their wines were excessively bitter, astringent and hard to drink. We noticed that if you take out too much alcohol (say less than 10%), you see another kind of harshness we called acid-based astringency. At first we assumed that wines would become more balanced as we approached 13%-14% alcohol on a preference bell curve.

Not so. After thousands of blending trials, we finally had to admit the obvious, the last thing we were expecting. Every wine has very discrete balance points we all can identify, where the astringency abates and the flavors are married and harmonious. The adjacent wines in such a series taste especially disharmonious -- 0.1% alcohol too high was hot and bitter, and 0.1% too low was harsh and sour.

In a recent experiment on a 2007 Cabernet Sauvignon, we found sweet spots at 13.7% and 14.2%, which 12 judges rated an average 6.0 and 6.1 respectively on a 10-point scale for astringency, while the blend of the two at 13.95% was rated 25% more astringent, with an average rating of 7.5 (confidence level very highly significant at α< 0.1%).

A key point is that this effect, which has nothing to do with phenolic content or structure, is very strongly shared. Something in the blend is perceived as on or off in the same way that if the piano is out of tune, everybody leaves the bar without even knowing why.

Here’s a fun experiment you can do at home: Open a nice big Cabernet Sauvignon, taste it and take note of the level of astringency. Boot up iTunes, put on the Doors’ “People Are Strange” and taste again. You’ll see the wine smooth out. Then put on “When The Saints Go Marchin’ In” (Louis Armstrong’s Golden Legends version will work). The wine will become almost undrinkably harsh. If you have a glass of White Zinfandel around, the effects will be the opposite.

What’s going on here? My notion is that wines, like music, carry emotional modalities. Cabernet likes dark, sinister music because it resonates with its nature. White Zinfandel likes silly, positive party music. When the wine’s spirit comes into conflict with its environment, the midbrain revokes access to its upper recesses and the stimulus, classified as noise, is sent to the brain’s primitive limbic recesses. Harshness ensues.

Whether or not that makes sense, this experiment dramatically demonstrates the importance of context for astringency perception. In addition, it shows how strongly shared is our sense of harmony, however divided we are as to individual preference. Holding simultaneously these disparate concepts in mind is a challenging exercise in mental gymnastics, good for building the philosophical muscle postmodern winemaking requires.

Clark Smith is winemaker for WineSmith, founder of the wine technology firm Vinovation. He lectures widely on an ancient yet innovative view of American winemaking. To comment on this column, e-mail

1. Gawel, R. (1998) Red wine astringency: a review. Australian Journal of Grape and Wine Research 4:74-95.
2. Adams, D., Harbertson, J. F., and Ried, M.S. (2004) Quantitative and Qualitative Differences in Red Wine Astringency. (
3. Noble, A.C. (1990) Bitterness and Astringency in Wine. In: ‘Bitterness in Foods and Beverages.’ Ed. R.L. Rouseff (Elsevier Science: Amsterdam) pp. 145-158.
4. Fischer, U., Boulton, R.B. and Noble, A.C. (1994) Physiological factors contributing to the variability of sensory assessments: Relationships between salivary flow rate and temporal perceptions of gustatory stimuli. Food Quality and Preference 5, 55-64.
6. Gawel, R., Oberholster, A. and Francis, L. (2000) A “Mouth-feel Wheel:” terminology for communicating the mouth-feel characteristics of red wine. Australian Journal of Grape and Wine Research 6: 203-207.
7. Food Chem. 73:423-432. 2001;
8. American Journal of Enology and Viticulture, March 1987; 38:69-77.
9. Smith, C. R. and Fugelsang, K. E. (2001) Winegrape Maturity Enhancement via Reverse Osmosis. Proceedings of the O.I.V. Groupe d’Expertes sur la Technologie du Vin, Paris, France.

Currently no comments posted for this article.