From Tangy to Tender: A Battle of Coagulants in Early Modern vs. Modern Tofu
Vinegar (Left) and GDL (Right) Tofu
Introduction:
Soybean has been a versatile vegetable that has been responsible for feeding millions of people from the day it was domesticated by Chinese farmers back in the 1100 BC (History of Soybeans 1). While the origins cannot be traced exactly, soybean seeds were founded in “archaeological sites connected to the Longshan and Yangshao cultures in China (dating to c. 3000–1500 BC and 5000–3000 BC respectively), those associated with the Chulmun culture in Korea (8500–1500 BC) and Japan’s Jōmon culture (10,500–300 BC)” (Thomas 28). Soybeans have been an important crop for these East Asian countries that fast forwards to the 21st century and there are distinct practices of how each of these countries utilized soybeans. For instance, for the Chinese, some signature soybean products that came out are stinky tofu, and Mapo tofu, focusing on the different textures tofu can give. For the Japanese, it is their fermented products such as miso, and natto shaping their culinary culture. As for the Koreans, it is their fermented paste such as gochujang and doenjang. This list illustrates the different varieties that each culture has cultivated throughout millennials. Now in modern times, as a society we have soy milk, tofu skin, dofu hua, bean curds; all the different delicious delicacies that can be made sweet, sour, and spicy.
Li Shizhen’s accreditation to Prince Liu An’s tofu Recipe:
Looking into how tofu making even happened, it’s important to investigate the origins starting with poets and Prince Liu An of the Han Dynasty (Needham et al. 303) where Li Shizhen a few centuries later credited the recipe to him. In addition to Prince Liu An’s initial discovery, a few dynasties later, fellow poets like Cheng Hsiung-Tuan wrote, “Turn the quern to let the jade liquid flow. Cook the milk and let the clear fluid glow.” (Needham et al. 303) These poetic instructions left a lot of open-ended interpretations by the readers especially “jade liquid” and “milk” in the second sentence. While the audience can interpret that the jade liquid is the milk, it’s very open-ended and often will lead to confusion. Pivoting away from poems, Li Shizhen was the first to provide step-by-step instructions compared to other poems at that time illustrating the tofu-making process with poetic verses. His recipe instructions were:
Soak the beans [in water].
Grind the beans [to give a bean milk].
Filter the milk [to remove coarse residues].
Cook the milk [to homogeneity].
Add yen lu (bittern), leaf of shan fan (mountain alum), or vinegar to curdle the milk.
Collect the curd.
But it is important to note that Li Shizhen can’t be a credible writer to trust as he had made claims that are often mythical in his book Compendium of Materia Medica. For example, according to the “historian of China Carla Nappi, ‘a man obsessed’ who filled his 52-volume pharmacological tome […] with just about everything […] ‘prescriptions for dragon bone, stories of corpse-eating demons and fire-pooping dogs…” (Thomas 21). There were a lot of wild claims that were made by Li Shizhen in his comprehensive book which also led to “the attribution of the invention of tofu to Liu An in Li’s Compendium [being] flimsy evidence, a corroboration made roughly 1,800 years after Liu An’s death…” (Thomas 21).
Research Question:
Comparing the modern tofu making to early modern Chinese tofu making that was fully documented in the Ming dynasty by Li Shizhen, the different agents that make the tofu curds coagulant are reflected in both recipes. By investigating how the texture, firmness, and other sensory properties [such as taste harmony] of tofu differ when prepared using traditional vinegar coagulation versus modern glucono delta lactone [GDL] coagulation, this experiment hopes to find answers on modern versus traditional preferences. We can further answer:
- Does vinegar-coagulated tofu result in a softer or firmer texture compared to GDL-coagulated tofu?
- Does vinegar-coagulated tofu retain a noticeable acidic taste compared to the neutral flavor of GDL-coagulated tofu?
- How do sensory evaluations (taste, mouthfeel) differ between traditionally made (vinegar) and commercially prevalent (GDL) tofu?
- Broader Question: What does the omission of detailed instructions on tofu making in Li Shizhen’s text suggest about the nature of knowledge in his time? Might this reflect an assumption that such processes were so widely understood that they required no explanation? Additionally, could this indicate the presence of tacit knowledge—practical expertise that was commonly practiced but not explicitly recorded?
Modern Recipe V.S Early Modern Recipe:
To test and attempt to answer some of these research questions, it was important to use a modern recipe that also offered a visualization aspect. Watching the visualization can lead to developing some tacit skills by mimicking since Li Shizhen’s recipe was super vague in description. Fellow YouTube creator Adam Witt made a video named Making Tofu from Scratch with Fresh Soybeans, and his instructions were:
Start by soaking your soybeans overnight. This will make them plump up and turn a brighter yellow.
Blend your soaked soybeans with water until a smooth puree forms. Pour this mixture into a medium saucepan.
Slowly bring the mixture to a simmer, skimming the foam from the top for cleaner milk and easier straining. Let it simmer for about 10-15 minutes.
Strain the mixture using a cheesecloth over a fine mesh strainer. Squeeze out as much of the liquid as you can. Set the leftover soy mash (also known as Okara) aside for future use.
You've now made homemade soy milk. Next, prepare your coagulant. For this recipe, we're using Nigari, diluted in water.
Warm the soy milk to around 170 degrees Fahrenheit (76 degrees Celsius), then introduce the Nigari mixture in two steps, stirring gently after each addition. Allow the soy milk to sit covered for 5 minutes after each addition.
Your soy milk should now coagulate and resemble a broken sauce. This is good and means your Nigari has done its job.
Line your tofu mold with a couple of layers of cheesecloth. Pour the coagulated soy milk into the mold, wrap it up in the cheesecloth, and place a weight on top. The weight and duration of pressing will determine the firmness of your tofu.
After about 30-40 minutes, remove the tofu from the mold and immerse it in a bowl of cold water for at least 30 minutes. This helps it tighten up and retain its shape.
Your homemade tofu is now ready to use immediately or can be stored submerged in water in the fridge for up to five or six days.
Adam’s recipe is much more detailed with descriptions of what the materials should look like after soaking or even adding the coagulant. The added visualization of his YouTube video was very helpful in running the experiment for both recipes since it gave a baseline of instructions. On the other hand, Li Shizhen’s recipe directions were vaguer, which makes one wonder if the people at that time already knew the fundamental processes of tofu-making, where recording the details wasn’t as important.
Ingredients/ Materials Needed:
Figure 1: Materials
· Soybeans
· Vinegar
· Cold Water
· Blender/ Food processor
· Cheese Cloth
· Bowls
· Ladle
· Scale
· Measuring Cups + Spoons
· Containers
· Sifter
· Tofu Mold
· Weights
· Glucono Delta Lactone (GDL)
Preparation/ Soaking the Soybeans:
Modern GDL Recipe Experiment, Challenges, Results:
Following Witt’s recipe was relatively easy since there was a visualization aspect at every step. For instance, Witt illustrated how soybeans should look after they’re soaked in water, the process of cooking the raw soybean milk, and the technique of adding the coagulant of choice into the heated soymilk mixture.
Some challenges that I faced while following Witt’s recipe were the size and wattage of my blender. My blender has a 250-watt power, while Witt’s Vitamix blender is more powerful, and he can do one batch all at once. Instead of trying to blend all 310 grams of soybeans in one go, I split the batch into 6 different blends with the ratio of 52 grams of soybeans to 1 cup of water. Splitting the mixture into 6 different blends, made the blending process very smooth and I was able to get a smooth homogeneous light-yellow color as seen in Figure 12.
Figure 15: Figure 1: Cooking the Raw Milk
After successfully blending the soybean and water mixture, the mixture was dumped into the steel pan without further separation all according to Witt’s instructions. There was a raw, sour smell from the mixture that was a little bit concerning but was mitigated after the mixture cooked through the 15-minute interval as seen in Figure 15. The mixture also was visually reduced in Figure 16 and the fresh soymilk also started to develop this film on the surface. When mixing the mixture with the ladle, the bottom of the pan seemed scorched/burned with the bean curd film developing. Towards the end, there was a burning scent in the kitchen due to this film on the bottom of the pan.
Figure 16: Figure 1: Milk Reduced
Witt’s method of sifting the mixture from the soybean clumps was to use the double strainer method where he put the cheesecloth on top of his deep strainer. This method was tried, but it was messy where the warm soymilk mixture leaked all over the bowl and overflowed on top of the cheesecloth. That method was abandoned, and the metal sifter was strictly used as seen in figure 18.
Witt uses Nigari, which is a salt solution, but GDL was substituted here. The ratio of GDL to the water mixture was 2 grams to 2 tablespoons of warm water mixed well. Witt reheated the cool soymilk in a pan up to 170 degrees, to a simmer before introducing the Nigari to start the coagulation. Similar steps were followed, but instead of introducing the coagulation agent into the mixture, it was poured into the clear bowl first. Then the introduction of the hot soymilk was added on top of the GDL mixture. This “pouring method” is also similar to Dou Hua making where instead of making tofu, this develops into tofu pudding (Zhu 6) as seen in Figure 23. Letting the mixture rest without disturbance resulted in this silky pudding-like solid that’s different from Witt’s coagulation mixture from his nigari addition.
From the solidified tofu mixture, a dessert was created with the addition of ginger syrup. The texture on top was very smooth and pudding-like, but the bottom (Figure 25) of the cube had this grainy texture that was not fun to eat. It left a sour taste in the mouth, and the ginger syrup didn’t cover that taste up either. The texture wasn’t homogenous like commercial tofu.
Figure 29: Unveiling of Tofu in Water
Putting the soft mixture into the metal pan with the cheesecloth was challenging because the cubes were soft and slippery. It was hard to move the mixture to one side of the pan because the tofu cubes were so delicate that cubes broke upon movement. As seen in Figure 26, on the side of the cheesecloth, the little bits of tofu cube remain can be seen. In addition, the soft tofu mixture made it easy to pack into a comprehensive square and the weights from the sparkling water can push the remaining liquid out and into the drainage system that was created. The cans sat on the tofu cubes for 30 minutes before unveiling the cheesecloth to reveal the block of tofu created in Figure 29. Thirty minutes was a baseline according to Witt, but if a firmer tofu block was preferred, more time could be added with weights. One of the challenges with adding more time to weight is the tofu block water retention. If too much water is pressed out, the texture and structural integrity of the tofu block can be altered. So, it was safer to follow the guidelines. After unveiling the tofu block, cold water was added to help tighten the remaining protein. This was left overnight in the fridge for further insurance of helping the firmness.
The process of taking out the chilled tofu was nerve-wracking due to breakage in the middle of the block. Producing tofu with GDL “…reduces whey separation. GDL also enables in-place coagulation…” (Kurrat and Perkins 8). Due to the higher water content that is exhibited with GDL as well as the in-place coagulation that was used with the “pouring” method, this might not allow the tofu cubes to combine as closely. Unlike other traditional coagulants like vinegar that allows that texture to fuse, GDL tofu cubes didn’t fuse together as well which illustrates the breakage after some movement. Looking closely at Figure 31, even cutting slices within the block led to some parts of the tofu falling apart into tofu crumbs. The resulting texture of GDL was less firm than store-bought tofu but firmer than silken tofu. The tofu had a beany taste and in terms of texture, the top of these slices’ cubes (Figure 32) was smooth pudding-like, while the bottom was grittier. That grittiness could be due to the GDL not mixing properly.
Early Modern Recipe Experiment, Challenges, Results:
Following Li Shizhen’s recipe was tougher than Witts because of the limited information regarding measurements of the beans, vinegar, and water. Li’s recipe also calls for a grinder to grind the beans into milk, which was difficult for a college student to obtain due to expense as well as space constraints. Additionally, Li’s recipe calls for filtering the milk first before cooking it, which was different from Witt’s recipe of cooking the mixture and then filtering it. Li’s first step of soaking the beans in water was straightforward but in terms of timing, led to some confusion. Li didn’t specify if the soybeans needed a full day to soak or if a couple of hours will make it soft. To solve this challenge, we opt to soak the beans for the same amount of time as Witt’s recipe.
Figure 34: Filtering Blended Soybeans
In the second step of grinding the beans, since a grinder wasn’t available, a blender was used as a substitution. We arrived at a homogenous raw mixture as seen in Figure 34. Li’s description of “bean milk” was confusing because his instructions didn’t explicitly call for adding some sort of liquid. To prevent damage to the blender, 1 cup of water was added to every 52 grams of soybeans. This was repeated 6 times to get the 310 grams of total soaked soybeans.
In Li’s third step of filtering the milk, we utilized the metal sifter again, as his instructions were straightforward. The smell that was coming out of the raw soymilk was not pleasant as it had a very earthy sour scent.
Li’s fourth step of cooking the milk until homogeneity was subjective since the milk already looked homogenous (Figure 37). An interesting indicator of whether if this mixture was done cooking was the raw sour smell that wafted through the air. Using that as an indicator of the eliminated sour smell and Witt’s recipe of letting the raw mixture cook for 15 minutes, it solved the raw smell. The milk tastes like regular soymilk.
Li’s step of adding the coagulant, he mentioned different products like yen lu and leaf of Shan fan which are uncommon household items. But his option of vinegar as an coagulant was accessible as there was 5% distilled vinegar in the kitchen. We also diluted this vinegar into 2 tablespoons of water to 1 tablespoon of vinegar ratio because we were afraid that it would make the tofu too vinegary. After letting the mixture simmer with the added vinegar coagulant, the curdles were very small (Figure 44). To speed up the process, one more tablespoon of vinegar was added to the mixture, and the curds immediately developed.
In the final step of collecting the curds, Li didn’t mention any details for how to pack or weighing the bean curds into a cohesive block. To mitigate this issue, we repurposed Witt’s instructions. The tofu was weighted down for 30 minutes. Compared to the GDL tofu, when packing it down, these curds were harder to pack together due to their bounciness. Even with the weights stacked on top of the collected tofu, some areas that were not covered by the weights bounced upwards, resulting in an uneven-looking tofu block (Figure 52). Cold water was also added to the block and was stored in the fridge overnight. After taking out the tofu block from the fridge, the texture was very crumbly. However, the cutting process was very smooth compared to the GDL tofu block. Bean crumbs weren’t spread out on the cutting board, and the vinegar tofu block was cut cleanly.
The tofu squares were very firm. When observed up close (Figure 54), we can see how the crumbles adhesive themselves to each other from the weights stacked on top. One would expect a very strong vinegary taste since 2 tablespoons of vinegar was added to the mixture, but surprisingly, the flavor was subtle. The fresh taste of tofu was very satisfactory added with the firm texture.
Takeaways:
After experimenting with both GDL and Vinegar as coagulants, Li’s recipe triumph in taste. His recipe not only has a firmer texture, but the slight vinegary taste adds a unique flavor to the tofu that the GDL tofu didn’t exhibit. The firmer texture of the ancient tofu can be attributed to the immediate separation of whey from the soymilk protein and the development of actual soy curds that were compressed into a block fully with weighted objects. According to Jungbunzlauer, “The unique texture of silken tofu is achieved by adding coagulants such as Glucono-delta-Lactone…, silken tofu production does not include a pressing step. As a result, more whey is bound in the product, giving the silken tofu its particular texture…” (Kurrat and Perkins 2). Interestingly, the GDL tofu in this trial, did go under the same 30-minute pressing and the resulting texture is less firm than the vinegar tofu block, but a little firmer than the store-bought silken tofu. The taste of the GDL tofu had that bean flavor, but it wasn’t as strong. There was also this gritty texture on the bottom of the tofu block that can be attributed to GDL not mixing as well because of the “pouring” method. In addition, the smell of both tofu blocks has a “stronger bean flavor compared to the more subtle flavor in packaged tofu, which can be attributed to its freshness…” (Won 1). The tofu that was made in these two experiments didn’t have any preservatives which most likely affected the taste and scent of the tofu as well. Overall, the making of tofu was a success. For both cubes, there wasn’t a weird taste besides the grittiness of the GDL tofu cube.
On a broader context, with the poems and even Li Shizhen’s recipe steps, the details were vague and poetic that they can cause confusion for people who would want to make tofu for the first time without any culinary background knowledge. It can be assumed that tofu making could have been such a widely known knowledge in the Han and Ming dynasties that recording it down step by step wouldn’t be useful. Tacit knowledge was also utilized in this experiment which partly confirms why the directions were vague. For instance, Li’s step of cooking the milk to homogeneity, tacit knowledge of previous cooking experience was needed to make this experiment functional. When the burning smell of the soymilk started, instinctively, the pan that contained the milk was lifted to mitigate that smell. Additionally, Li Shizhen’s Compendium of Materia Medica was more focused on medicinal and natural knowledge. Li Shizhen’s priority was more on explaining the health benefits, nature, and uses of products rather than step-by-step guidelines. For instance, Li Shizhen shared that for people who had drank too much alcohol, his suggestion was to have thinly sliced hot tofu and stick it all over the body, since tofu is branded as a ‘cooling food’ based on Chinese traditional medicine. (Thomas 21). All in all, Li’s purpose was not to provide specific recipes but simply inform his readers on nature, medicine, and influence Chinese traditional medicine in his own way.
Works Cited:
“History of Soybeans.” North Carolina Soybeans, 23 Feb. 2022, ncsoy.org/media-resources/history-of-soybeans/.
Kurrat, Johanna, and Lena Perkins. “Glucono-Delta-Lactone in Silken Tofu.” Jungbunzlauer, 7 July 2020, www.jungbunzlauer.com/fileadmin/content/_PDF/PRINT_PROJECTS/Article_facts/JBL_AR_Glucono-delta-Lactone_in_silken_tofu_2020-077_HR.pdf.
Li, Shizhen. Compendium of Materia Medica Book 2 Comp. by Li, Shizhen. Transl. and Ann. by Luo, Xiwen. Foreign Languages Press, 2003.
Needham, Joseph, et al. Science and Civilisation in China. Volume 6, Biology and Biologicial Technology. Part VI, Medicine. Cambridge University Press, 2004.
Thomas, Russell. Tofu: A Culinary History. Reaktion Books, 2024.
Witt, Adam. “Making Tofu from Scratch with Fresh Soybeans | From Scratch.” YouTube, 21 Oct. 2020, youtu.be/_grQxw2H2Ds?si=DdIp2rj1ezDhA7el.
Won, Jay. “Making Tofu: A Staple in East Asia.” Kitchen, Studio, Factory: Making in East Asia, Kitchen, Studio, Factory: Making in East Asia, 5 May 2022, www.kitchenstudiofactory.com/rework-project/making-tofu.
Zhu, Maggie. “The Ultimate Guide to Making Douhua (Tofu Pudding).” Omnivore’s Cookbook, 24 Dec. 2024, omnivorescookbook.com/douhua-tofu-pudding/.
