Made in a Beaker, Not on a Farm

A Short History of Industrial Food: How the American Kitchen Changed

Before the twentieth century, the American kitchen ran on fats that came directly from the farm or the slaughterhouse — lard rendered from pork fat, butter churned from cow's milk, tallow scraped from beef, and occasionally coconut or palm oil imported for specialty uses. These were whole or minimally processed substances with long culinary traditions; a hog raised and butchered on a farm yielded the same fat that had fried cornbread and greased cast iron for generations. The kitchen, in other words, was downstream of the pasture and the orchard. That relationship began to fracture in the opening decades of the 1900s, when chemical engineering and industrial scale made it possible — and commercially irresistible — to manufacture fats rather than grow them.

The pivot point arrived in 1911, when Procter and Gamble introduced Crisco, an all-vegetable shortening produced by hydrogenation — a process that forces hydrogen gas through liquid cottonseed oil under heat and pressure, converting it into a solid, shelf-stable white fat. Cottonseed oil had previously been regarded largely as an industrial byproduct — a waste stream from cotton ginning operations — with little established place in the American diet. Hydrogenation transformed that waste stream into something that looked and behaved remarkably like the lard it was positioned to replace. The resulting product was formulated in a facility, assembled through chemistry, and designed from the outset for stability and margin — not grown on a farm and rendered in a kitchen.

Those industrial origins ran deeper than ginning waste. Before anyone cooked with it, cottonseed oil was burned in lamps in place of whale oil, boiled into soap and candles, and used to grease machinery — an industrial input, not a pantry staple. When lard grew expensive in the late 1880s, meatpackers began quietly cutting it with the cheap oil, an adulteration that a congressional inquiry eventually forced onto labels as “lard compound.” Hydrogenation offered a cleaner second act: the same oil, hardened, whitened, and renamed, could finally enter the kitchen through the front door instead of the back.

Procter and Gamble's marketing campaign for Crisco was, by the standards of its era, extraordinarily sophisticated. The company distributed free cookbooks containing recipes that called specifically for Crisco, effectively rewriting the assumed vocabulary of American home cooking. Salesmen visited cooking schools and domestic science programs, which were themselves relatively new institutions organized around the idea that cooking was a technical discipline amenable to scientific improvement. The pitch was wrapped in the language of purity and modernity: Crisco was white, odorless, and consistent in a way that lard — an animal product varying by season and breed — could never quite be. The industrial product was framed not as a substitute but as an upgrade.

The timing was precise. The Progressive Era brought with it a genuine and warranted anxiety about adulteration in the food supply — Upton Sinclair's 1906 novel about the meatpacking industry had crystallized public unease, and the Pure Food and Drug Act had passed that same year. Into that atmosphere of distrust toward animal products and unregulated processing, a manufactured shortening packaged in a sealed tin and presented as scientifically pure found receptive audiences. The irony is dense in retrospect: a product assembled by industrial chemistry benefited from public skepticism of industrial handling of animal fats. The messaging worked because it exploited a real anxiety and redirected it toward a manufactured solution.

Over the following decades, the logic that built Crisco was applied across a widening range of pantry staples. Vegetable oils pressed from soybeans, corn, and cottonseed — crops grown at industrial scale in the American Midwest — were refined, bleached, and deodorized through multi-step chemical processes that altered or removed color, flavor, and many of the compounds present in the raw seed. These refined liquid oils then moved into bottles sold at grocery chains, into the vats of commercial fryers, and into the formulations of packaged crackers, margarines, and baked goods. The century's middle decades saw soybean oil in particular climb to extraordinary volume as government agricultural programs incentivized soybean production and food manufacturers found a versatile, inexpensive input. A fat that had been absent from the American diet before the twentieth century became, by mid-century and beyond, among the most widely consumed added fats in the country.

What the arc of industrial fat reveals is a broader pattern in how the American food system was reorganized across the twentieth century: the kitchen became a downstream market for what facilities could produce at scale and price, rather than a place where farm outputs were prepared into meals. Whole ingredients were displaced not because they were found deficient for culinary purposes, but because manufactured substitutes could be produced more cheaply, transported without spoilage, and promoted with the language of science and modernity. The shift from lard to Crisco was not a single dramatic rupture but an early, legible example of a system-wide transformation — formulated in a facility, shipped to a shelf, and sold as progress.

How a Seed Becomes an Oil: The Industrial Refining Process

The journey from field to bottle begins not with a press but with a solvent. Modern commodity oil production — for soybean, canola, corn, and cottonseed — relies on hexane extraction, a petroleum-derived chemical used to pull fat from the crushed seed meal with extraordinary efficiency. The seeds are first cleaned, cracked, and flaked to maximize surface area, then flooded with liquid hexane in a continuous extractor. The hexane-oil mixture is then heated so the solvent evaporates off and can be recovered and recirculated — but trace residues can remain in the crude oil at this stage. The result is a dark, pungent, wax-cloudy liquid that bears no resemblance to the clear, neutral product on a grocery shelf.

What comes out of the extractor next enters a multi-stage chemical sequence called refining, degumming, and neutralization. Degumming removes phospholipids and other gum-forming compounds by treating the crude oil with water or phosphoric acid; these gums, left in, would cause cloudiness and interfere with later processing. Refining — sometimes called alkali refining — follows, treating the oil with sodium hydroxide (lye) to remove free fatty acids, which are then separated as a soap-like byproduct called soapstock. What remains is a cleaner oil, but still far from the pale, odorless fat that industrial food manufacturing demands.

Bleaching is the third transformation. The degummed, neutralized oil is mixed with bleaching earth — a fine absorbent clay — which binds to pigments, oxidation byproducts, and residual trace metals. The mixture is filtered under vacuum, stripping the oil of color compounds including chlorophyll and carotenoids. The oil emerges visibly lighter, but it is not yet the shelf-stable, flavoring-neutral fat that formulators require. Food science literature has documented that the heat and chemical exposure at this stage can produce minor trans-fatty acid isomers and oxidation byproducts as process artifacts — an acknowledged consequence of industrial processing conditions.

The final and most dramatic stage is deodorization — the step that completes the transformation from agricultural raw material to engineering ingredient. The oil is exposed to high-pressure steam at elevated temperatures — industry literature commonly cites a range in the vicinity of 230–260°C — for an extended period in a near-vacuum environment. This strips out volatile compounds responsible for the oil's natural flavor, odor, and any remaining off-notes. It is, in effect, a distillation: everything that made the oil taste or smell like a seed is removed. What exits the deodorizer is a fat without identity — colorless, odorless, flavorless, and engineered to disappear into any formulation.

The economic and functional logic of this process is inseparable from its scale. The entire refining sequence is designed to produce a cheap, undifferentiated, shelf-stable fat that behaves predictably across thousands of different manufactured products — frying oils, margarines, snack coatings, baked goods, and salad dressings. The invisibility is a feature, not a side effect: food formulators need a neutral fat that contributes calories and texture without competing with the flavor system engineered into the product. This is precisely what the process delivers — a commodity input as interchangeable and specification-controlled as an industrial chemical, produced from grain in a large-scale facility, then shipped in bulk tankers to factories that formulate it into what eventually reaches a plate.

It is worth appreciating what is lost or altered in the transformation. Cold-pressed or expeller-pressed oils — produced by mechanical pressure without solvents — retain natural antioxidants, phospholipids, and minor compounds present in the original seed. The high-heat industrial process destroys or removes most of these. What remains after hexane extraction, bleaching, and steam deodorization is a highly purified fat fraction, stripped of context — a molecule optimized for price, clarity, and shelf life. Formulated from grain in a facility, not pressed on a farm: the distinction is both literal and structural.

THE THOUSANDFOLD SHIFT

At the dawn of the twentieth century, the American kitchen ran almost entirely on animal fats — lard rendered from pork, butter churned from cream, tallow trimmed from beef. These were slow, seasonal, and expensive to produce. Then, in 1911, Procter and Gamble introduced Crisco, a partially hydrogenated cottonseed oil sold in a white tin that looked deliberately like lard. It was formulated in a facility, not rendered on a farm, and it cost less than the thing it was designed to replace. That was the opening move of a century-long substitution that would transform what sits inside nearly every packaged food on the American market.

The industrial logic behind seed oils was, in its own terms, elegant. Cotton was already a dominant crop, and the seed — pressed out during fiber production — was largely a waste product. Chemists found they could extract its oil, deodorize it, bleach it, and hydrogenate it into something spreadable. The process required nothing that grew slowly or grazed on pasture; it required a refinery. As the technology scaled, it was applied to soybean, corn, sunflower, safflower, and canola, each of which could be planted across millions of industrial acres and processed at volumes no dairy herd or hog operation could match. The result was a class of fats that were cheap at scale, reliably neutral in flavor, and stable on a shelf for months — three properties that made them ideal raw material for a packaged-food industry that was itself just beginning to industrialize.

The scale of the shift over the following decades is not a matter of scientific dispute: it is a matter of agricultural and economic record. Per-capita consumption of soybean oil alone in the United States rose from near zero at the start of the twentieth century to dramatically higher levels by the century's end — a transformation without precedent in the history of the American fat supply. Animal fats did not disappear, but their share of the fat supply shrank steadily as vegetable oils moved from a curiosity to a commodity to a near-universal ingredient. By the time a child born in the 1980s was eating school lunch, seed oils were present in the salad dressing, the bread, the crackers, the dipping sauce, and the cookies — reflecting that oil had become the most economical way to deliver fat, texture, and shelf life across a packaged-food supply chain.

What made seed oils so useful to manufacturers was precisely what made them unlike anything found in a traditional kitchen. A bottle of extra-virgin olive oil, pressed from a whole fruit, is far more perishable than the refined alternatives it competes against; a jug of refined soybean oil, stripped of its naturally occurring pigments and trace compounds through deodorization and bleaching, resists oxidation long enough to survive a transcontinental supply chain. That stability is not accidental — it is engineered into the product. The refining process removes much of what would otherwise limit shelf life, leaving behind a fat that is, in a meaningful chemical sense, a more industrial product than the seed it came from. Formulated in a facility. Made to last.

What followed was a feedback loop between agriculture, manufacturing, and dietary guidance that proved self-reinforcing. As seed-oil production scaled, prices fell. As prices fell, food manufacturers substituted them into more products. As more products contained them, per-capita consumption rose. Government dietary guidelines issued in the latter half of the twentieth century, responding to concerns about saturated fat, frequently encouraged the substitution of vegetable oils for animal fats — guidance that had the effect, whatever its intent, of accelerating a shift already well underway for industrial reasons. The system optimized for volume, cost, and stability, and the American diet was reshaped around what that system produced.

Whether this dramatic shift in the composition of the fat supply has been beneficial, neutral, or harmful to human health is a question the scientific literature has not settled. Researchers continue to debate the relative effects of different fatty-acid profiles, the significance of the refining process, and the degree to which seed oils as actually consumed in ultra-processed foods can be studied in isolation from everything else those foods contain. What is not debated is the scale of the transformation itself, or its industrial origins. A fat that barely existed in the American diet in 1900 now shows up in the overwhelming majority of packaged foods. It was not cultivated into existence by farmers; it was manufactured into dominance by a food system that needed something cheap, stable, and invisible — assembled in a plant, not grown on a farm.

Cottonseed oil greased machines before it ever greased your dinner. The recipe didn't change much — just the label.

Grease for Engines, Sold as Dinner

Rapeseed oil spent most of its history nowhere near a kitchen. For centuries it was a lamp fuel and a lubricant, and when World War II arrived it became strategically critical: naval engineers found rapeseed oil valuable for lubricating the wet metal surfaces of marine engines in ways that other available lubricants could not reliably replicate. Countries stockpiled it. Navies ran on it. It was, without ambiguity, an industrial product.

The reason it stayed out of the food supply was chemistry. Rapeseed is naturally high in erucic acid, a fatty acid that the body handles poorly in large amounts. The oil was approved for machines, not mouths, and that is where it sat — until Canadian plant breeders in the early 1970s began a selective breeding program specifically designed to engineer the erucic acid out of the plant and produce a version palatable enough to sell at the grocery store.

They succeeded. The new low-acid variety was registered around 1978 under a name designed to put as much distance as possible between the product and its industrial past: canola — commonly spelled out as “Canadian Oil, Low Acid” — built to sound clean and neutral. The old name was a liability. The new name was a marketing decision. Nothing about the underlying plant species changed — Brassica napus, the same crop that had been lubricating engines for decades — only the branding and the acid profile.

The playbook was not new. Cottonseed oil ran the same route a century earlier: once a byproduct of the cotton industry with little use at the table, it was processed, hydrogenated, and sold as a cooking product once the technology caught up with the commercial incentive. Canola is cottonseed with better public relations and a Canadian origin story. The pantry absorbed another industrial oil, and most people filling their grocery carts never once thought to ask where it came from.

The Trans-Fat Cautionary Tale

The story of trans fat is, at its core, a story about what happens when industrial chemistry moves faster than biological understanding. In the early twentieth century, food manufacturers faced a practical problem: liquid vegetable oils went rancid quickly, were inconvenient to ship, and behaved unpredictably at high cooking temperatures. The solution arrived in the form of hydrogenation — a process in which hydrogen gas is forced through liquid oil under heat and pressure in the presence of a metal catalyst, rearranging the fat molecules into a semi-solid state that was shelf-stable, cheap, and easy to work with. The result was partially hydrogenated oil, and it was, from an industrial standpoint, a near-perfect ingredient. It could be produced at scale from commodity crops, stored for months without spoiling, and pressed into nearly any food that needed fat. What it was not was an ingredient whose long-term profile had been rigorously established before it reached the supply chain at scale — a gap that regulators would spend decades filling in.

Crisco, introduced in 1911, was among the earliest mass-market products built on this chemistry — a white, solid shortening sold as a modern, scientific alternative to lard. The pitch was credibility through novelty: a fat built in a factory rather than rendered from an animal, pure and consistent in a way that nature could not guarantee. Food manufacturers adopted partially hydrogenated oils enthusiastically across the decades that followed. By the second half of the twentieth century, they were embedded in the supply chain of packaged crackers, margarine, fast-food frying oils, baked goods, and snack foods. Artificial trans fat — the molecular byproduct of the hydrogenation process, produced when some of those rearranged fat molecules took a geometrically unnatural form — was present across a wide range of everyday foods. Because no labeling requirement existed at the time, consumers had no routine way to know how much they were consuming or from which products.

For years, the presence of trans fat in food was not merely accepted but, in some quarters, actively promoted. When public health concern over saturated fat from animal sources grew through the 1970s and 1980s, partially hydrogenated vegetable oils were sometimes positioned as the more virtuous substitute. The irony would become plain only later. Evidence on trans fat was accumulating slowly in the research literature, and the industrial food system had a decades-long head start. The ingredient was already baked — literally — into thousands of products, into restaurant fryers, into the economic infrastructure of food manufacturing. Changing course would mean not a quiet reformulation but a systemic overhaul.

The regulatory turning point came in stages. The FDA required that trans fat be listed on the Nutrition Facts label beginning in 2006 — an acknowledgment that consumers deserved to know what they were eating, and a signal that the agency's posture toward the ingredient was shifting. Then, in 2013, the FDA made a preliminary determination that partially hydrogenated oils — identified as the primary source of artificial trans fat in the food supply — were no longer Generally Recognized As Safe, a designation known as GRAS. In 2015, the FDA finalized that determination, giving manufacturers a compliance deadline that effectively required the removal of PHOs from the food supply. The phrase "no longer generally recognized as safe" is precise regulatory language, and it means something specific: the scientific consensus that had once supported the ingredient's use had, in the FDA's judgment, been overtaken by evidence sufficient to warrant removal.

What the trans-fat episode demonstrates is not that the food industry acted with malice, but that the industrial food system operates on a timeline that is structurally mismatched with the timeline of biological research and regulatory review. An ingredient can be adopted at massive scale — embedded in supply chains, consumer habits, and decades of product formulations — before the long-term picture comes into focus. By the time the FDA acted, partially hydrogenated oils had been a staple of the American food supply for roughly a century. The reversal required, and received, a formal regulatory determination that an ingredient once presumed safe was no longer so. It is, in the clearest possible terms, a documented precedent: engineered in a facility, adopted at scale, reversed decades later.

Painted Food: The Synthetic Color Palette No Farm Ever Grew

Open a bag of neon-orange cheese puffs or a bottle of electric-red maraschino cherries and you are looking at paint — not pigment that occurs anywhere in the natural world, but synthesized colorants derived from petroleum chemistry and certified by the FDA as safe for human consumption. These dyes exist for one documented purpose: cosmetic. They do nothing for taste, nutrition, texture, or preservation. They are applied because the industrial food system discovered, across decades of commercial practice, that color is a purchase driver — and that a product manufactured on a production line rather than grown on a farm often needs visual repair before it can reach the shelf.

The roster of certified synthetic dyes is designated with names like Red No. 40, Yellow No. 5, and Blue No. 1. The FDA's certification process attests that each batch meets chemical purity standards; it does not certify that the dye confers any benefit to the person eating it. The color is strictly cosmetic, a coat of paint applied to products whose natural appearance was destroyed or degraded somewhere in the refining, extruding, or reconstituting process that turned raw ingredients into packaged goods.

Red Dye No. 3 occupied the certified palette for decades — showing up in maraschino cherries, certain candies, and a range of processed foods — before the FDA revoked its authorization for use in food. The agency had findings from animal studies, completed by 1990, showing that the dye caused thyroid tumors in male rats at high doses — through a hormonal mechanism the FDA has said does not occur in humans — but the revocation did not follow immediately. The agency acted decades later, and the mechanism that compelled the action was the Delaney Clause, a provision of federal law that prohibits the approval of any food additive shown to cause cancer in humans or animals — regardless of whether the agency believes the risk carries over to people. In other words: the FDA's own position was that it posed no demonstrated human risk, yet a 1958 law still forced its hand, decades after the data landed. The gap between the animal-study findings and the eventual regulatory action is a documented feature of how slowly the system can move.

State-level legislation in California moved ahead of federal action. AB 418, signed into law in 2023, banned four additives — brominated vegetable oil, potassium bromate, propyl paraben, and Red Dye No. 3 — from food products sold in the state, making California the first U.S. state to act legislatively on a group of additives still permitted under federal law at the time. The following year, AB 2316 addressed the school food environment specifically, prohibiting six synthetic dyes — including Blue No. 1, Blue No. 2, Green No. 3, Red No. 40, Yellow No. 5, and Yellow No. 6 — from being served in California public school meals. The legislative record for both bills documented that the dyes in question are already banned or restricted in the European Union and the United Kingdom, jurisdictions that had applied a precautionary standard where the United States had not.

What makes synthetic dyes a particularly clear example of formulated food is that they are, in the most literal sense, unnecessary. Whole foods are already colored — by anthocyanins in blueberries, carotenoids in carrots, chlorophyll in spinach. The need for petroleum-derived substitutes arises only when a product has been processed so aggressively that its natural color is gone, or when a product is being engineered to look like something it is not. The electric orange of a snack chip, the fire-engine red of a fruit punch that contains no fruit, the uniform golden-yellow of a cheese product: these are corrections applied after the fact to food that has already been stripped of the qualities that would have made the color honest. Made in a beaker, not on a farm — and then painted to look otherwise.

The regulatory activity now underway — state-level legislation in California and the belated federal revocation of Red Dye No. 3 — marks an institutional reckoning with a part of the food supply that was always cosmetic rather than nutritional. Whether the federal framework converges with what the science and several international regulatory bodies have already acted on remains an open question; the Red Dye No. 3 timeline, measured in decades from animal data to revocation, illustrates how extended that process can be. What is not in dispute is the underlying fact: certified color dyes were put together in a processing plant, approved for a cosmetic function, and added to the food supply at industrial scale for generations, well after regulators in other countries had drawn a different line.

THE FLAVOR IN THE PACKAGE

When a food manufacturer lists 'natural flavors' near the bottom of an ingredient panel, it reads like a minor footnote — a finishing touch, a background note. In practice, it can be one of the most chemically complex entries on the label. The U.S. Food and Drug Administration defines 'natural flavor' as any flavoring derived from a natural source — a fruit, vegetable, herb, spice, meat, seafood, dairy product, or fermentation product — but the definition says nothing about how many individual compounds may be present, what solvents or carriers those compounds arrive in, or in what concentrations they were concentrated, extracted, or recombined. A single 'natural flavor' entry can represent a formulation that contains dozens of distinct molecular components.

The flavor industry that supplies these formulations is a specialized and largely invisible branch of the food supply chain. Flavor houses — dedicated contract manufacturers — develop proprietary blends for food brands, and those blends are legally protected as trade secrets under FDA rules. Because the flavoring is classified as a single ingredient for labeling purposes, the brand selling the product is not required to disclose what the flavor actually contains, and in many cases the brand itself may not always have full visibility into the supplier's formula. This is not fraud; it is the system operating exactly as designed, a design that dates back to regulatory frameworks built before ultra-processed food existed at its current scale.

The reason 'natural flavors' occupy such a prominent role in formulated foods connects directly to what industrial processing does to raw ingredients. High-heat extrusion, hydrogenation, refining, and extended shelf-life packaging strip or degrade the volatile aromatic compounds that give whole food its sensory character. A corn chip made from whole dried corn tastes like corn because the corn's own volatiles survived. A snack made from refined corn starch, corn syrup solids, and a thin slick of oil tastes like nothing until a flavor formulation is introduced to replace what was engineered out. The flavor, in other words, is not a bonus — it is structural, a reconstruction of sensory experience that the manufacturing process itself destroyed.

Flavor chemists work with a palette that can include hundreds of individual aroma compounds, many of them identical in molecular structure to compounds found in whole foods. The word 'natural' in the regulatory definition refers to origin — the compound was derived at some point from a biological source — not to the process by which it was concentrated, isolated, or recombined. Benzaldehyde, the compound responsible for almond and cherry notes, can be derived from peach pits and labeled natural; it is also identical in every chemically meaningful way to the same compound synthesized in a laboratory. The distinction between 'natural' and 'artificial' flavor, in molecular terms, is often a distinction without a practical difference.

There is no settled scientific consensus that consuming these flavor compounds at typical dietary doses causes harm, and the FDA maintains a GRAS (Generally Recognized As Safe) framework under which many flavor ingredients are evaluated. What is documented, and not in dispute, is the structural opacity: a consumer reading 'natural flavors' cannot determine from that label entry what they are ingesting, how many compounds are present, or what ancillary carriers and solvents accompany the primary flavoring. For people managing food allergies, religious dietary requirements, or specific ingredient avoidances, this opacity has real practical consequences — a reality that consumer advocates and researchers have noted in the ongoing public conversation about ingredient transparency.

What 'natural flavors' ultimately represents, in the context of formulated food, is the closing act of reconstruction. The seed oils provide the fat structure; the refined starches and sugars provide the caloric base; the emulsifiers and stabilizers hold the matrix together; and the flavor system — assembled in a flavor house, shipped as a concentrate, and disclosed in two words — tells your palate a story about the food that the food itself can no longer tell. It is taste without origin, sensation without source. The product on the shelf is not grown, fermented, or cured into flavor. It is formulated into it, ingredient by ingredient, compound by compound, in a facility most consumers will never see and whose contents a label was never designed to reveal.

They engineered a fat designed to pass through you largely unabsorbed, then sold it to you as the healthy choice.

The GRAS Loophole: How Additives Enter the Food Supply Without a Regulator's Blessing

Most Americans assume that if something is in their food, someone in Washington signed off on it. The reality is more unsettling. The U.S. food additive system contains a provision — the Generally Recognized as Safe designation, universally known as GRAS — that allows a manufacturer to determine, on its own, that a substance it wants to put in food is safe. Not safe pending review. Not safe upon approval. Safe by the company's own conclusion, with no mandatory obligation to tell the FDA it made that call. The mechanism is not a loophole in the conspiratorial sense; it is written into the framework. That is precisely what makes it worth understanding.

The foundation was laid in the Food Additives Amendment of 1958, which required pre-market approval for new food additives. Congress carved out an exception for substances that experts already recognized as safe at the time — common ingredients like salt, vinegar, and baking powder that had a long history of use. The logic was sound: the FDA did not need to formally review table salt. But the exception was defined broadly enough that over the decades it stretched far beyond pantry staples. The GRAS category became a door that could be opened for entirely novel substances, including compounds formulated in industrial facilities, not grown on a farm, provided a manufacturer could assemble a panel of credentialed experts to agree on safety.

The self-determination pathway works like this. A company developing a new ingredient commissions a safety review — typically conducted by scientists it selects and pays. If that panel concludes the substance is generally recognized as safe among qualified experts, the company may begin using it in food. Critically, under this interpretation of the rules, the company is not required to notify the FDA, submit data for independent review, or wait for any regulatory response. The decision lives inside the company. The FDA has operated a voluntary notification program, where manufacturers may inform the agency and receive a 'no questions' letter, but the operative word is voluntary. A substance can enter the food supply and reach store shelves before any regulator has seen the supporting data.

The consequence is a structural information gap. The FDA cannot evaluate what it does not know exists. Reformulations, novel processing aids, and newly synthesized flavor compounds can move from a manufacturer's facility into packaged food products through this pathway without generating a public record. This is not a fringe outcome or a rare exploit. It is a documented feature of how the modern processed food supply is built — ingredient by ingredient, panel by panel, each step technically within the rules, the cumulative picture not systematically visible to the agency nominally responsible for food safety. The system was designed for a simpler era; it is operating in an era of industrially engineered food at continental scale.

None of this means every GRAS substance is harmful. Many have been reviewed independently, used for decades, and face no serious scientific challenge. The problem is architectural: the process optimized for efficiency over transparency, and efficiency won. When a substance's safety case is built by the party with a financial interest in a favorable conclusion, reviewed by experts that party selected, and filed in a drawer rather than submitted to regulators, the public has limited means to interrogate that conclusion. Regulatory agencies in other countries require pre-market notification or approval for the same substances that enter U.S. food via GRAS self-determination, a disparity that reflects a genuine policy choice about where the burden of proof should sit.

The GRAS pathway matters because it is the on-ramp. Understanding how substances get into the food supply does not require conspiracy; it requires reading the mechanism. Formulated ingredients — compounds engineered for texture, shelf stability, palatability, or cost reduction — can reach your plate through a process that bypasses the public review system most people believe is in place. The gap between that assumption and the documented reality is the story. Not the individual ingredients on any given label, but the system that decided who decides whether those ingredients belong there at all.

Engineered to Pass Through You

What if a food company could sell you fat — actual fat — and simultaneously advertise that you're not absorbing it? Not a trick of language. Not a rounding error on a nutrition label. A molecule, designed from scratch, whose entire selling point is that your body cannot process it.

That is Olestra. Approved by the FDA in 1996, Olestra is a synthetic fat built by esterifying sucrose with six to eight fatty acids — a structure so large and geometrically locked that human digestive enzymes cannot break it down. It passes through you. That was the goal. Frito-Lay sold it in WOW chips under the brand name Olean, and the pitch was straightforward: all the crunch, none of the fat calories, because none of the fat is actually absorbed. The FDA signed off. Then it required the company to print a warning on every bag, in plain language, that the product “may cause abdominal cramping and loose stools” and that it “inhibits the absorption of some vitamins and other nutrients.” Vitamins A, D, E, and K — the fat-soluble ones — were stripped out along with the engineered fat that was never supposed to be there in the first place, so the manufacturer added them back. The warning label stood from 1996 to 2003, when the FDA removed the requirement after reviewing further data — but the mechanism never changed. The molecule still does exactly what it was designed to do.

The industry learned from Olestra. Not to stop, but to be quieter about it. The modern successor is EPG — esterified propoxylated glycerol — a modified oil engineered so that digestive enzymes cannot break it down. Because so little of it is absorbed, its caloric contribution is a fraction of conventional fat — by the maker's own accounting, around 0.7 calories per gram against the 9 in ordinary fat. It cleared the FDA not through the full approval process but via the GRAS pathway — Generally Recognized as Safe, a category that allows manufacturers to make their own safety determination. No mandatory warning label. No federally required disclosure on the front of the package telling you that a chemist designed this molecule specifically so your gut would reject it.

Here is the thing worth sitting with: this is not food that happens to be processed. This is not a shortcut taken during manufacturing, not a preservative added for shelf life, not a dye chosen for aesthetics. This is a substance built at the molecular level with indigestibility as the product feature. The entire value proposition — the reason it exists, the reason it was approved, the reason it is in food at all — is that your body cannot use it. It was designed to fool your digestion. Then it was marketed to you as the healthy choice.

It Can't Even Call Itself Ice Cream

Stand in the freezer aisle and read the cartons closely. A surprising number of the things that look like ice cream do not, legally, get to call themselves that. The label says “frozen dairy dessert.” That phrase is not a marketing flourish. It is a disclosure required by law — a quiet admission that the product in your hand failed to clear the bar for the real thing.

The FDA reserves the word “ice cream” for a product that meets an actual standard of identity (21 CFR 135.110): at least 10 percent milkfat, a minimum weight of milk solids, and a minimum heft per gallon. Real cream, in real amounts, frozen. Miss that threshold and you forfeit the name. What gets made instead — lighter on dairy fat, frequently propped up with stabilizers, gums, and other engineered ingredients to fake the body and mouthfeel that the missing cream used to provide — is a “frozen dairy dessert.” Same shape, same pint, same freezer. Different food.

It is a small tell, and once you see it you see it everywhere. “Frozen dairy dessert” where you expected ice cream. “Pasteurized process cheese food” where you expected cheese. Each of these is a legal flag, mandated precisely because the product fell short of the standard it is imitating and had to be relabeled around the gap. The name on the box is downgraded because the contents were.

That is the whole story of the modern supermarket compressed into a label. When a food has to invent a softer name for itself, it is usually because it stopped being the food it is pretending to be. Real ice cream is cream, milk, sugar, and cold. Everything past that is engineering — and the fine print, if you bother to read it, will quietly confess as much.

Engineered to Be Eaten: How Ultra-Processed Foods Became Industrial Formulations

Walk into any American supermarket and count the products with more than five ingredients you could not reproduce in a home kitchen. The number is not incidental — it is the point. Ultra-processed foods are not cooked; they are formulated. They begin not with a crop or an animal but with a specification: a target texture, a target color, a target shelf life, a target cost-per-serving. Ingredients are selected the way an engineer selects components — for their functional properties first, their origin second or not at all. The result is a product that travels through the food system less like a perishable good and more like a manufactured durable.

The industrial process that produces these foods is as far from a kitchen as a semiconductor fab is from a workshop. Raw agricultural commodities — corn, soy, wheat — are first stripped of the fiber, bran, and micronutrients that make them perishable, yielding neutral starches, isolated proteins, and refined oils that are shelf-stable, odorless, and effectively blank. Those refined intermediates are then recombined with emulsifiers, stabilizers, humectants, and flavor compounds into a finished product whose sensory profile — its taste, its crunch, its color, its mouthfeel — was dialed in at a development bench before a single unit was manufactured. The raw material is agriculture; the finished product is chemistry.

Central to this process is a concept food scientists call the bliss point: the precise concentration of sweetness, saltiness, or fat at which measured consumer liking is maximized. The term entered public discourse through the work of food industry researchers who, by the 1970s and 1980s, were running systematic sensory tests — essentially optimization experiments — to find the stimulus intensity that produced peak preference scores. The bliss point is not a metaphor. It is a measurable, reproducible target on a dose-response curve, and hitting it reliably is a core competency of modern food product development. A product engineered to land precisely on that target is, in the technical vocabulary of the field, hyperpalatable — calibrated to elicit maximum eating enthusiasm. Researchers studying hyperpalatability have noted that the flavor combinations it exploits — fat paired with sugar, fat paired with salt — appear together in nature only rarely, and never at the concentrations found in formulated products. The hypothesis, advanced in peer-reviewed nutrition literature and still actively debated, is that the human sensory system evolved in an environment where those combinations reliably signaled energy-dense food worth consuming aggressively, and that formulated products can present those signals at intensities that have no natural analog. Whether and how that dynamic drives overconsumption at a population level is genuinely contested and ongoing — researchers disagree about mechanisms, about study designs, about how much to extrapolate from controlled laboratory conditions — but the engineering intent is not contested at all. The products are designed to be eaten with enthusiasm, and the industry measures success partly by whether they are.

The scale at which this now operates is the most striking fact. Researchers analyzing nationally representative dietary survey data have estimated that ultra-processed foods account for more than half of daily caloric intake among U.S. adults — one widely cited analysis, published in BMJ Open, put the figure at roughly 57 percent of energy intake, though estimates vary depending on the classification system used. Whatever the precise share, the directional finding is consistent across multiple data sets: the majority of the energy Americans consume on an average day arrives not from whole or minimally processed ingredients but from products assembled in a manufacturing facility from refined fractions, additives, and synthesized flavor systems. A generation ago that share was far smaller; the shift maps closely onto the industrialization of the food supply and the sustained decline in the real cost of manufactured food relative to fresh ingredients. What was once a supplement to the diet — the packaged snack, the shelf-stable convenience item — has become the diet's structural center.

The running ingredient list of a typical ultra-processed product captures the logic in miniature. Where a home cook might use butter, a formulated product may use an interesterified oil that delivers a similar mouthfeel at a fraction of the cost and with a shelf life measured in months rather than weeks. (Partially hydrogenated oils, once the dominant tool for this purpose, were largely phased out of the U.S. food supply after the FDA determined in 2015 that they were no longer Generally Recognized As Safe and required their removal.) Where a baker uses eggs for emulsification, a food technologist uses lecithin or mono- and diglycerides — molecules that perform the same structural function but are produced through industrial fat-processing rather than cracked from a shell. Natural color fades and varies across batches; certified synthetic dyes do not. The cumulative effect is a product that resembles food in every sensory dimension — it looks right, it smells right, it feels right in the mouth — but whose component list reads less like a recipe than like a bill of materials. Formulated in a lab, not grown on a farm.

GRAS doesn't mean a regulator checked it. It means the company that makes it decided, on its own, that it's fine.

WHERE WE HAVE ENDED UP

Walk the center aisles of any American supermarket and the scale of the shift becomes impossible to ignore. The CDC has reported that roughly 2 in 5 U.S. adults now meet the clinical definition of obesity, with severe obesity — a body-mass index at or above 40 — now present at rates that are meaningfully higher than mid-twentieth-century baselines. These numbers represent a genuine epidemiological change, a shift that researchers, clinicians, and public health officials have spent decades trying to explain. No single explanation has achieved scientific consensus, and the honest answer is that the cause is almost certainly not singular at all.

At the same time, the American food supply has undergone its own documented transformation. Nutrition researchers who track dietary patterns have estimated, using data from national dietary surveillance and frameworks such as the NOVA classification — a system developed by food scientists to categorize industrial formulations that incorporate additives, flavoring agents, emulsifiers, and other substances not typically found in a home kitchen — that ultra-processed products now account for more than half of the calories consumed by the average American adult, and a substantially higher share for children and adolescents. A supply chain that once moved relatively whole ingredients from farm to table now moves, in very large part, engineered products formulated in facilities, assembled from fractionated components, and designed to stay shelf-stable for months or years.

These two facts — a historically high prevalence of obesity and a food supply dominated by industrially formulated products — exist together in the same historical window, and researchers, journalists, and public health advocates often note that co-occurrence. What the science does not yet cleanly establish is the degree to which one is connected to the other, in which directions any relationship runs, or how to partition responsibility among the many variables that changed simultaneously: labor patterns, sleep, activity levels, portion norms, economic stress, the built environment, the collapse of home cooking, and the composition of the food itself. The honest scientific position, as of now, is that the relationship is real and deeply studied but not fully resolved.

What is not contested is the texture of the change in the food supply itself. Ingredients that did not exist in commercial food a century ago — partially hydrogenated oils introduced in the early twentieth century and later phased out after regulatory action, synthesized colorants, high-fructose corn syrup at industrial scale, dozens of shelf-life extenders — became ordinary components of the American diet within living memory. The FDA has taken documented regulatory action on some of these substances over the years, including the revocation of the generally recognized as safe status for partially hydrogenated oils, a process completed in 2018. The food was, in a traceable and documented sense, increasingly made in a facility rather than grown on a farm.

The people eating this food did not choose the transition in any granular sense — they chose from what was available, affordable, and aggressively marketed, within economic constraints that made whole-food cooking increasingly difficult for large portions of the population. Epidemiologists observe that the burden of obesity falls unevenly, concentrated in lower-income households and communities with less reliable access to fresh food, a pattern that resists any explanation centered purely on individual choice. A food environment is not a neutral backdrop; it is itself a product of decisions made across corporate, agricultural policy, and retail economics over decades.

None of this is a clean story. The science of nutrition is genuinely hard — randomized controlled trials of diet are expensive, difficult to blind, and run over short time horizons relative to the chronic conditions in question, while long-term epidemiological data is confounded by the sheer number of things that change at once in a society. Researchers disagree about the relative contributions of specific macronutrients, specific additives, specific processing methods, and lifestyle factors entirely outside of diet. What the data does show, clearly and repeatedly, is that the American population is eating a diet unlike anything its ancestors ate, that rates of obesity and metabolic conditions are high and still rising, and that these two facts share a timeline. What to do with that shared timeline is the question the science is still working to answer.

The Slow Reversal

The regulatory machinery moves slowly, but it does move. In January 2025, the FDA revoked authorization for Red Dye No. 3 — a synthetic colorant that had been used for decades to make maraschino cherries and certain candies glow a vivid red — after studies in lab rats prompted the agency to act under the Delaney Clause, a federal provision that prohibits approving additives shown to cause cancer in animals. A year earlier, in 2024, the FDA revoked authorization for brominated vegetable oil, or BVO, a compound that had been used as an emulsifier in citrus-flavored drinks for decades. And in 2022, the European Union banned titanium dioxide as a food additive — a white pigment used to make candies, frosting, and sauces appear brighter and more uniform — citing safety concerns the agency concluded could no longer be dismissed. Three additives, three regulators, three separate decisions spanning three years. Not a revolution. A slow, reluctant acknowledgment that the beaker-built food supply requires ongoing audit.

What these reversals share is instructive. In each case, the additive was not a nutrient. It was not there to feed you. It was there to optimize appearance, texture, or stability — to make a product look more appealing on a shelf, stay emulsified in a bottle, or maintain a uniform white opacity across millions of units. These are engineering problems, solved with industrial chemistry, then cleared for consumption under regulatory frameworks built on the assumption that low doses in isolation pose no concern. The regulatory retreats suggest that assumption is being stress-tested. They do not settle the science. They mark the places where confidence ran out.

The deeper problem is that for every additive walked back, others remain in wide use — some approved decades ago under evidentiary standards that would look thin by today's benchmarks, some that have never been independently reviewed at all. The GRAS designation — Generally Recognized as Safe — allows manufacturers to self-affirm the safety of new ingredients without mandatory FDA review. The result is a system in which the ingredient list on a packaged food represents not a fully audited safety dossier, but a rolling historical record of what has not yet been challenged. That is not a scandal. It is a structural feature of how industrial food has been governed. But it is worth understanding clearly before you decide how much weight to put on any single label claim.

This is the honest case for cooking from whole food — not a cure for anything, not a rejection of modernity, but a practical exit ramp from a system that was built around shelf life and sensory appeal rather than around you. When you start with an ingredient that grew or grazed rather than one synthesized and stabilized, the burden of proof flips. You are not waiting for a regulator to eventually revoke something. You are working from a shorter, legible list by default. That is what clean means here — not a marketing category, not a certification, but a structural constraint: food grown on a farm, not formulated in a lab.

The reversals of Red Dye No. 3, BVO, and titanium dioxide did not make the food supply clean. They are waypoints in an ongoing audit of a system that was assembled faster than it could be fully understood. The productive response is not outrage at any particular ingredient or agency — it is to build eating habits that are structurally insulated from whatever the next revision turns out to be. Cook real food. Read short ingredient lists. Treat ultra-processing itself — the extraction, recombination, and stabilization cycle that turns a crop into a shelf-stable product — as the relevant unit of concern, not any single additive. The system will keep revising. You do not have to wait for it.

What Clean Actually Means

The word "clean" gets thrown around so loosely in food marketing that it has nearly lost its meaning — but the underlying concept is durable and simple: whole food is food that has not been formulated. An apple is an apple. A chicken thigh is a chicken thigh. A handful of dried lentils is exactly what it says on the bag. The moment a product requires a seven-line ingredient deck to explain itself, something has been added, removed, or industrially transformed along the way. That transformation is not inherently evil, but it is worth understanding — because the formulation process, by design, optimizes for shelf stability, cost, and palatability, and those goals are not always the same as yours.

Reading a label is a learnable skill, and the core of it is surprisingly mechanical: scan for ingredients that could not plausibly exist in a home kitchen. A bottle of olive oil should contain olive oil. A jar of almond butter should contain almonds and, if you like, salt. When a product lists a partially hydrogenated fat, a petroleum-derived color, or a multi-syllable emulsifier you would need a chemistry degree to pronounce, you are looking at something compounded in a lab rather than prepared from food. That is not a moral judgment about the manufacturer — it is just a description of what the ingredient is and where it came from.

This is where a deterministic ingredient check becomes genuinely useful, because human memory and willpower are finite and label-reading at the grocery store is fatiguing by design. A rule-based system does not get tired. It does not make exceptions because a package has a picture of a farm on it or uses the word "natural" in the brand name. It applies the same standard to every ingredient on every label, every single time. The question it answers is not "does this feel clean?" but "does every ingredient in this product pass a documented standard for whole-food sourcing?" That is a different question, and a much harder one to game.

Cooking from whole ingredients is the most reliable version of that standard, because the ingredient list is you — what you chose to buy and combine. A meal built from a protein, a fat, vegetables, and a starch does not require a label at all. The skill involved is modest: most meals that qualify as genuinely whole-food take less than an hour to prepare and use fewer than ten ingredients. The barrier is not really technique. It is the accumulated habit of reaching for the packaged option first — a habit that the modern industrial food system, with its decades of marketing investment and shelf-space engineering, has had considerable time and resources to reinforce.

None of this requires perfectionism. A pragmatic clean-eating framework acknowledges that processed foods exist on a spectrum — minimally processed canned tomatoes are not the same category of thing as a shelf-stable snack cake engineered to never go stale. The useful question is not "is this 100% unprocessed?" but "do I recognize every ingredient as something that was once grown or raised?" When the answer is yes, you are in whole-food territory. When the answer requires a reference database, you are in formulated territory — and knowing which side of that line you are on is already most of the work.