by Dwight Furrow
Much philosophical writing about food has included discussions of whether and why food can be a serious aesthetic object, in some cases aspiring to the level of art. These questions often turn on whether we create mental representations of flavors and textures that are as orderly and precise as the representations we form of visual objects.
The claim that we do not form such ordered, mental representations is central to the view that food and drink cannot be serious aesthetic objects or works of art. The reason is that genuine aesthetic experience requires the apprehension of form or structure. In the absence of structured representations there is no form to apprehend. (See Part 1 of this essay for a more detailed account of mental representations of aroma.)
I want to argue that the skeptics have a point although they draw the wrong inference from it. Our flavor experiences do, in fact, rest on weak representations. They lack the stable, detailed, clearly identifiable objects that vision typically yields. However, these weak representations allow for the apprehension of a different kind of structure, what I shall call a nomadic distribution (or continuously changing structure), and it is in fact this nomadic distribution of flavor that enables the aesthetic experiences characteristic of our engagement with food. The main point to draw from this is that cuisine is about transformation, one thing becoming another, and it is ill served by theoretical perspectives that rely on mental representations with firm identity conditions.
The explanations for why flavor and aroma are weakly representational fall into three categories—flavors are ephemeral, ambiguous, and difficult to remember with precision.
Ephemerality
The dishes we consume at meals exist for only a few minutes and are destroyed as we consume them. Thus, appreciation must happen quickly. This is an obstacle to fully appreciating dishes that are complex or unfamiliar. Although in principle we can repeat the experience of a dish until we understand it, diners typically don’t return to the same restaurant or continuously choose the same dish in order to grasp its form or structure. Chefs, on the other hand, do have the opportunity to understand the structure of the dishes they create, but an aesthetic practice in which only producers grasp aesthetic form is peculiar. Habits surrounding home cooking also don’t facilitate the kind of repetition or analysis akin to grasping the meaning of a painting or patterns in a musical work. This is largely because most home cooks lack the time, budget, and technical skills to create dishes that are replete with formal aesthetic character.
Ambiguity
The second reason for the weakness of mental representations of flavors is that they are often ambiguous and difficult to identify and describe. Flavors form complex mixtures of interpenetrating qualities, each flavor shaping the others in the mixture. These interwoven attributes differ by degrees, depending on the precise mix of flavor molecules, their amounts, and the heat applied to them. Thus, they lack clear identities in that they cannot be sorted into precisely defined categories. Furthermore, they are subject to non-linear effects that significantly influence what we taste and how we react. Non-linearity in this context means that small changes in a flavor or texture can produce large, cascading effects on other flavors which interact in ways that defy simple ordering or progression.
To use a simple example, raw tomato has a distinctive flavor although that flavor differs depending on the variety. The flavor of lightly cooked tomato is clearly related to raw tomato but nevertheless markedly different. Cooked tomatoes that have been exposed to sufficient dry heat over time will lose their moisture and begin to caramelize, among other chemical transformations, gaining dimensions that the raw or lightly cooked versions lack. Small differences in how much heat is used or how long it is applied make a significant difference in how a dish tastes. Add garlic and we have another qualitative shift, the result of which is highly sensitive to minor differences in the amount or strength of the garlic. Add ground meat and cook for several hours as in Bolognese sauce and the transformation is even more profound.
Bolognese sauce is a far cry from a raw tomato. Has the tomato lost its identity via these transformations? There is no right answer to this question because there is no clear line of demarcation that marks off the identity of an ingredient or its characteristic flavor. Everything depends on which flavors are combined and how they are handled. More complex sauces such as a mole that may include 20 or more ingredients will exhibit a variety of these transformations, each affecting how the other ingredients are perceived. Of course, tomato sauces are familiar components of many dishes across a variety of cuisines, and that repetition facilitates the formation of stable, predictable mental representations. But unfamiliar, innovative, or exceedingly complex dishes—the kind we encounter in contexts where dishes are presented as distinctively aesthetic objects—are challenging because of the ambiguity of flavors and the non-linear effects of their combination.
Furthermore, as we consume a dish, especially one with some complexity, the evolution of flavors and textures in the mouth from diverse ingredients produces differences across multiple points rather than a single, absolute flavor. For instance, a carne asada taco combines salsa, avocado sauce, and marinated beef wrapped in a corn tortilla. Each component evolves with varying levels of intensity and qualitative change as it is consumed, the flavors appearing as distinct yet also combined in a mixture with continuously shifting degrees of relative salience.
These non-linear transformations, which are often unpredictable when new combinations are formed, mean that cooking is relentlessly experimental. Add ¼ teaspoon of salt to a bowl of soup and the flavors brighten. And another ¼ teaspoon and the soup becomes discernibly salty, add another and it may be inedible even though the percentage of salt in the dish is low compared to other ingredients. However, as any chef will report, how much salt a dish needs is not a matter of a precise value correlated with a given volume but a matter of intensity, degrees of variation regulated by other ingredients and seasonings. The precise balance of flavors in a dish can never be achieved through invariant rules. Flavors vary by degrees of intensity rather than by clear quantitative measurements and this requires that we grasp flavor experiences not via fixed properties but as a continuous variable process sensitive to contextual factors. Tasting, not measurement, is the means of discerning balance. Degrees of variation rather than clearly delineated properties are pervasive in cuisine.
In summary, flavor is a product of mutually intersecting gradient fields rather than discreet flavor entities. As a result, flavors lack constancy because they are deeply influenced by fluctuating contexts. This contextualism can be extended to the environmental conditions in which we eat. We know that visual appeal, atmosphere and seasonality, the music being played, and the company at the table also influence what we taste.
The non-linear character of flavor has an additional implication that helps explain why grasping detailed flavor relations is a challenge, especially when dishes are unfamiliar. Because infinitesimal differences can have substantial phenomenological effects, it is difficult to explain these effects without knowing the details of how the dish was designed and produced. Much that goes on in a dish in terms of how flavors are influencing each other is phenomenologically inaccessible. Again, chefs may have this knowledge, especially if they know the chemistry of flavor, but diners for the most part won’t have access to it without pursuing specialized knowledge. (Lack of phenomenological access is also true of the substantial unconscious processing of taste and flavor signals in the brain, a complication I will leave aside for now.)
Memory
Finally, flavors and the patterns they form are difficult to remember. Although we are reasonably good at remembering generic aromas and flavors, we have difficulty forming precise memories of subtle variations or unfamiliar mixtures of flavors, the kind which are relevant to aesthetic appreciation. The difficulty we have in identifying even familiar aromas by name exacerbates the weakness of our flavor memories.
As I pointed out in the earlier essay, some of these difficulties can be mitigated to a degree through the kind of disciplined pedagogy we find in professional wine tasting (or beer, cheese, or coffee tasting.) But this mastery is tenuous and subject to repeated failure and is in any case unavailable for the more complex domain of cuisine.
This ambiguity and lack of clear definition among flavor relations has been used to cast doubt on whether some food preparations might be works of art. However, this critique of the aesthetic potential of food and drink assumes that aesthetic appreciation is anchored in accurate perceptual representations. There are reasons to think that the capacity to form accurate, detailed representations of food and drink is only one dimension of experiencing them aesthetically and when emphasized excessively misses a more important dimension—the ability to track and gain pleasure from variation.
This ability to track variation is facilitated by weak representations.
Why is the appreciation of variation essential to gastronomy? The answer lies in how our sensory mechanisms work, especially regarding the phenomenon of sensory adaptation. Sensory adaptation is the process by which our sensory receptors become less sensitive to constant or unchanging stimuli over time. It allows us to focus on changes in our environment rather than background sensations, enhancing our ability to detect new or important stimuli. When you walk into a room with a strong odor, you might initially find it overwhelming. However, after a few minutes, you hardly notice it. This is because olfactory receptors reduce the strength of the signal to your brain. If you eat something very sweet, your sensitivity to sweetness can decrease temporarily, making other sweet things taste less intense immediately after. Sensory adaptation enables us to filter out repetitive stimuli and respond more effectively to variation in our environment.
Sensory adaptation has a variety of implications for cuisine. Dishes must be designed with sufficient variation to avoid a feeling of satiation. Menus and meals must be varied enough so we maintain sufficient stimulation to avoid boredom and maximize pleasure. Food cultures and traditions must be updated with new dishes and ingredients over time if their inhabitants are to maintain their appreciation of food.
Of course, this need for variation to block sensory adaptation runs up against another feature of our food consumption. We are creatures of habit. We have fond memories of Grandma’s meat loaf and crave its repetition, habitually eat the same foods out of convenience or inattention, and return again and again to favorite dishes? As important as it is to block sensory adaptation, we are also biologically wired to want to put in our bodies only foods with which we are familiar. Taste and aroma are the last line of defense. They warn us when we ingest something dangerous as well as indicating that what we are eating will enhance nutrition. Bitterness warns us of foods that contain toxins; sourness and “off” aromas warn of food that has spoiled; sweetness indicates the presence of an energy source; umami signals the presence of protein; saltiness indicates the presence of an essential mineral. We enjoy familiar foods because they have been vetted by experience as something safe and nutritious. Familiar foods are comforting and resonate with our feelings for the people we share those foods with and the food traditions they constitute.
Thus, we are creatures of habit but only up to the point where adaptation sets in when variation then becomes essential. Our appreciation of food is governed by both the need for familiarity and the need for variation and it is the tension between these conflicting needs that provides the foundation for the aesthetics of cuisine. The importance of variation is obvious in our contemporary, increasingly global, food culture where we routinely sample foods from outside our cultures of origin, and demand constant innovation in our Michelin-ranked temples of gastronomy. But the history of the ancient spice trade as well as the movement of staple ingredients across the globe long before modern technology facilitated such transfers indicate that variation has always been important when people have a choice in the matter. Tomatoes were introduced to Italy in the 15th century, potatoes introduced to India in the 17th century, curry rice was introduced to Japanese culture in the 19th century among many other examples, and each was assimilated to provide variation in the diets of people in tradition-bound food cultures.
Given this interest in variation in the foods we eat, we can understand why tastes and aromas are only weakly representational. Weak representations facilitate the recognition and assimilation of variation. Taste, like other forms of sensory experience, depends on successful pattern recognition. Yet the respective mechanisms of pattern recognition must be appropriate to the sensory data generated by each sensory modality. I have been arguing that the sensory data we get through taste and smell is of a different sort than the sensory data we get from vision. Vision apprehends a world of relatively stable objects with clearly articulated boundaries; taste apprehends a world of unstable, continuously changing mixtures in which degrees of variation and subtle almost infinitesimal differences can fundamentally change the experience.
In order to track variation in flavor, we need to constantly update preference patterns and improve our capacity to recognize complex yet unfamiliar patterns in the foods we eat. If the cognitive capacities we deploy to understand taste were built on the kind of precise conceptual categories and accurate, detailed memories that characterize a substantial portion of our visual experience, pattern recognition would routinely fail and would give too much weight to past experience. With stronger representations, updating preference patterns and cognitive recognitional capacities would take longer and require more resources. It should be noted that professional wine tasters must guard against this. Relying too much on entrenched ideas about what something should taste like, thus allowing expectations to override what one is tasting, is the bane of wine tasters, and the best learn not to rely too much on top-down processing of sensory data.
Thus, the fact that we form relatively weak representations of taste objects compared to visual objects is in keeping with the character of taste and aroma. What we appreciate when we enjoy cuisine is variation constrained by some degree of familiarity and that appreciation depends on having fluid categories and relatively weak memories that are more easily updated. Our weak representations of food and drink don’t inhibit our ability to recognize form and structure. Rather they enable us to recognize a different kind of structure, a nomadic organization that undergoes continuous modification.
Thus, an account of the aesthetics of food must rely on perceptual pattern recognition appropriate to the sensory modalities of taste and olfaction and not import forms of pattern recognition more appropriate for vision.