Fermentation technology as a driver of human brain expansion

Brain tissue is metabolically expensive. Consequently, the evolution of humans’ large brains must have occurred via concomitant shifts in energy expenditure and intake. Proposed mechanisms include dietary shifts such as cooking. Importantly, though, any new food source must have been exploitable by hominids with brains a third the size of modern humans’. Here, we propose the initial metabolic trigger of hominid brain expansion was the consumption of externally fermented foods. We define “external fermentation” as occurring outside the body, as opposed to the internal fermentation in the gut. External fermentation could increase the bioavailability of macro- and micronutrients while reducing digestive energy expenditure and is supported by the relative reduction of the human colon. We discuss the explanatory power of our hypothesis and survey external fermentation practices across human cultures to demonstrate its viability across a range of environments and food sources. We close with suggestions for empirical tests.

environments, temperatures, and food sources. We close with suggestions for empirical tests.

I. Introduction: The Problem of Hominin Brain Expansion
Human brains are notable for their large size. Over the course of 2 million years of evolution, the human brain has tripled in volume. Australopiths possessed brain volumes that were roughly the size of our closest living ape relatives, chimpanzees and bonobos (Pan troglodytes and Pan paniscus) (Holloway 1970;Tobias 1963;Dart and Salmons 1925). With the appearance of Homo, brain expansion in the human lineage began to accelerate, and continued through to the emergence of H. sapiens and H.
neanderthalensis. Although we have much information on the timeline and extent to which the human brain has expanded in our evolution, the mechanisms which drove this expansion are more difficult to determine. Several theories have been proposed, briefly summarized below.

Metabolic constraints on brain evolution
The Expensive Tissue Hypothesis (Aiello and Wheeler 1995) argues that the expansion of brain size in the hominin lineage required the reallocation of resources from the digestive system. In this view, the limiting factor for brain expansion is the availability of caloric resources, because brain tissue is metabolically expensive compared to most other tissue. Mutations leading to increased brain size, though they might support more adaptive behavior by the organism, could not actually be adaptive if they carried with them an increased risk of starvation. A reduction in the amount of gut tissue, which has metabolic needs similar to brain tissue, would free up the calories that would otherwise be used to support and maintain digestion and permit its reallocation to supporting the brain. Supporting this model is the fact that in addition to having relatively large brains, the size of the human gastrointestinal tract is 60% of that expected for a primate of our size (Aiello and Wheeler, 1995). However, because gut tissue is itself responsible for extracting nutrients from food, mutations leading to reduced gut size could not be adaptive without a prior shift to a more energy-dense, easy-to-digest food source. The Expensive Tissue Hypothesis has generated extensive empirical research, including studies that support this model (Kaufmann et al., 2003;Kotrschal et al., 2011;Tsuboi et al., 2015;Jin et al., 2015;Liao et al., 2016) as well as several that did not find a direct, causal tradeoff between metabolic investment in different tissue types, or instead support a more complex relationship with other metabolic investments van Schaik, 2005, 2006;Liu et al., 2014;Kotrschal et al., 2015). In particular, in a study of over 100 mammalian species, a consistent inverse relationship between gut size and brain size was not observed (Navarette et al., 2011). In a revised and elaborated model, Isler and van Schaik (2014) emphasize a complex system of tradeoffs between fitness benefits and multiple energetic costs including development, reproduction, digestion, and locomotion, where the cognitive benefits of a larger brain can only produce an increase in net fitness if the corresponding energetic costs are accounted for; notably, dietary changes are one proposed mechanism that could contribute to this.
This key point -that brain size increases are evolutionarily limited by metabolic constraints --has led to a search for possible changes in diet during the period when encephalization quotients began to depart from earlier anthropoids. Some proposed changes include increased meat consumption via hunting or scavenging, a specialization for the consumption of starchy underground tubers, or the development of cooking technology using fire.

Proposed dietary specializations of early human ancestors
Increased meat-eating has been argued to have been central to human evolution (Speth 1989;Katharine Milton 1999). Aiello and Wheeler (Aiello and Wheeler 1995) have further proposed that meat-eating is a plausible source of the extra calories needed to allow for brain expansion, and analysis of gut morphology in humans suggests it may be adapted to an intermediate diet with aspects of both frugivory and carnivory (Mann 2000). The archaeological record also supports the importance of meat eating in human ancestors, with fossil evidence of butchery in early Homo (Semaw 2000;de Heinzelin et al. 1999). It is certainly inarguable that modern human diets frequently involve more meat consumption than our anthropoid relatives, and data from modern hunter-gatherers support this view; however, some authors (e.g., (Cordain et al. 2002) argue that evidence for human hunting appears later in human evolution -in the Middle to Late Paleolithic. Another possibility is that meat was acquired by other means. addition to being labor-intensive to unearth, wild foraged tubers have as little as ¼ of the caloric density reported by Vincent (1985), even after cooking.
Another possibility is that the modifications to food through cooking provided the necessary additional calories and nutrients to support a reduction of gut and increase in encephalization (Wrangham et al. 1999). The hypothesis has been extended to encompass others. For example, cooked tubers have been proposed as an important component of the "cooked foods" diet (O'connell, Hawkes, and Blurton Jones 1999;Wrangham et al. 1999;Hatley and Kappelman 1980) and it has been suggested that scavenged carcasses were cooked to mitigate microbiological contamination (Smith et al. 2015). The trend of reduction of molar size in hominin evolution, suggested to be an adaptation from moving from tougher to softer foods (McHenry and Coffing 2000), fits well with this hypothesis (Zink, Lieberman, and Lucas 2014). The benefits of cookingincrease in bioavailability of calories, easier mechanical digestion (especially chewing), and the lowering of energy requirements for digestion -are undoubtable. However, there is a lack of archaeological evidence for the usage of fire by australopiths and early hominins; the earliest date for the evidence of fire by hominins is frequently cited at 0.5 mya by H. erectus during the Middle Pleistocene (James et al. 1989). Evidence for fire mastery in the Lower Pleistocene (Goren-Inbar et al. 2004) still puts this behavior well after the initial emergence of H. erectus, which is well after selection for brain expansion put hominins on a different course than the Pan lineage. It is almost certainly the case that the actual origins of human-controlled fire predate its oldest surviving archaeological evidence. However, more importantly, mastery of fire technology requires individuals to have the cognitive capacity to plan, create, maintain, and use fire effectively; this seems a tall order for an organism with a brain-to-body ratio not much exceeding that of modern nonhuman apes. This suggests that we should continue to search for other mechanisms that could have kicked off our ancestors' initial encephalization.

A New Hypothesis: External Fermentation
If we are to explain the dietary changes that supported brain expansion and gut reduction in the hominin lineage, we will need to identify strategies that were accessible by individuals with brains that were roughly the size of a chimpanzee's. Here, we outline a novel hypothesis, the External Fermentation Hypothesis (Figure 1). Central to this hypothesis is the realization that the gut is itself a machine for internal fermentation: digestion is accomplished via the the endogenous microbiome. Culturally-transmitted food handling practices which promoted the externalization of this functionality to the extra-somatic environment could have offloaded energetic requirements from the body and freed up the surplus energy budget for brain expansion.
In this paper, we begin with a mechanistic discussion on how external fermentation provides adaptive benefits: it increases macronutrient absorption; it increases the bioavailability of micronutrients, some of which are essential for brain development and function; it supports internal fermentation by the endogenous microbiome; and it provides additional immune benefits. Following this, we present evidence that external fermentation specifically addressed the expensive tissue problem: the reduction in human gut size is attributable mainly to reduction in the colon, which is the primary site of internal fermentation; furthermore, humans receive a surprisingly low amount of their calories from short-chain fatty acids (SCFAs), which are the products of colon fermentation on carbohydrates. Next, we consider the plausibility and explanatory power of the External Fermentation Hypothesis compared to other hypotheses.

II. Fermentation plays an important role in digestion
Fermentation is the breakdown of organic compounds by enzymes into alcohol, acids, or both. When discussed in the context of human metabolism and nutrition, the enzymatic activity typically originates from bacteria, yeasts or both, and transforms starches, sugars, and proteins into alcohol and/or acids. Rather than relying on the microorganisms living inside an animal's gut to ferment macronutrients, external fermentation is carried out by organisms living wild in the environment or on the surface of the organic material itself. Like internal, or intestinal, fermentation, external fermentation increases the bioavailability of ingested nutrients.
Digestion is the process of mechanically and enzymatically breaking down organic food matter into macronutrients small enough for absorption through the intestinal barrier and into the bloodstream. Any foodstuffs not broken down by enzymes, bile or other digestive chemicals pass through the upper gastrointestinal tract unabsorbed, offering the body no nutritional value. The digestion of fibrous, starchy vegetable matter requires a specialized digestive system with modifications that support fermentation. In ruminant animals, this is achieved through additional stomachs -these species are known as foregut fermenters. The hindgut fermenters, which include humans and other primates, as well as non-ruminant mammals -have evolved a large colon, large cecum, or both. A large colon and/or cecum means a large amount of surface area for absorption, but it also means a large amount of internal fermentation.
While both the large and small intestine contain active, symbiotic bacteria, the small intestine contains approximately one million bacteria per mL while the colon contains up to one trillion bacteria per mL (Gibson and Rastall 2004;Sender, Fuchs, and Milo 2016;Whitman, Coleman, and Wiebe 1998). Combined with a longer transit time than the small intestine (approximately 1-4 hours versus 18-39 hours), this means the action within the colon is focused on bacteria-driven fermentation. Although previously it was thought that in humans, the large intestine did little more than resorb water, there is a new focus on the significance of colon for human health, including immune responsivity

Fermentation promotes macronutrient absorption
Fermentation within the gut increases the body's capacity to absorb macronutrients beyond the normal function of the upper gastrointestinal tract. Fermented soluble fiber provides an average of 2 cal/g, an additional 50% to the 4 cal/g available from digestible starch and sugars. This energy is only available via the salvaging of otherwise undigested fiber through internal fermentation by gut microbes (World Health Organization 1997;Popovich et al. 1997);CFIA 2017;WHO/UN 1998). Notably, humans purposefully ferment feed for livestock (sileage) in order to increase its digestibility and caloric value. Like starches and sugars, fibers are polysaccharide structures made up of bound glucose molecules and other small carbohydrates.
Originating primarily in the cell walls of plants, fibers such as cellulose and pectin are resistant to hydrolyzation by human digestive enzymes and therefore pass through the small intestine unbroken (Messer et al. 2017;Nelms and Sucher 2015;Vanderhoof 1998). Once in the colon, these fibers are fermented by enzymes from gut flora, and the now-available sugars are fermented into acid/base conjugates. These are then further degraded by secondary microorganisms into short chain fatty acids (SCFAs) (Cummings and Macfarlane 1991;Bik et al. 2017;Battcock et al. 1998). Microbial fermentation of carbohydrates into SCFAs is estimated to contribute 2-10% of total dietary energy in humans (McNeil 1984;Livesey 1995;McBurney 1994). This is small compared to other mammals, which typically derive from 16% to over 80% of maintenance energy from the production of SCFAs in the gut (see Table 1).
These fermentation products have important biological functions. More than 80% of produced SCFAs take the form of either butyrate, proprionate, or acetate (Bik 2017). Butyrate is the preferred energy source for the cells making up the intestinal wall of the colon, and feeds the rapidly reproducing colonocytes while also producing Vitamin K and a variety of B vitamins for circulation (Vanderhoof 1998; Messer et al.

Fermentation promotes micronutrient absorption
Fermentation is also critical for the absorption of vitamins and minerals. One way this can occur is via actual synthesis of vitamins by bacteria. In the colon, vitamin K2 is Another mechanism by which internal and external fermentation increase bioavailablity of micronutrients is through the breakdown of anti-nutritional factors (ANFs). ANFs are compounds found in staple cereals, grains, seeds, legumes and tubers that bind essential nutrients, preventing their absorption in the body.  (Bassiri and Nahapetian 1977), yet sufficient absorption of these is critical for life (e.g., Lopez et al. 2016;Rerksuppaphol and Rerksuppaphol 2018;DiNicolantonio et al. 2018). Interestingly, humans, unlike rodents, produce little phytase in their small intestine (Iqbal, Lewis, and Cooper 1994). The bioavailability of minerals is therefore greatly reduced in humans despite their abundance in raw material. Lactobacillus bacteria-driven fermentation is an alternative to phytase --by lowering the pH, it provides a favorable environment for both bacterial and endogenous phytase within the plant material to hydrolyze the binding phytate and release the bound minerals (Humer and Schedle 2016;Katz 2016). Oxalate can also be degraded through Lactobacillus fermentations, either externally or internally (Bik et al. 2017;Wadamori, Vanhanen, and Savage 2014). Of note, degradation of phytate by external fermentation has been shown to be more effective than heat treatment or cooking due to the decreased phytase bioactivity at a temperature above 80°C (Gupta, Gangoliya, and Singh 2015; Mahgoub and Elhag 1998).
Another group of ANFs includes phenolic compounds like tannins that bind to proteins and enzymes, lowering their bioavailability (Nikmaram et al. 2017). Tannins also lower the overall digestibility of amino acids, minerals, and other macronutrients (Nikmaram et al. 2017 (Glander 1982), (Garber 1987)). Primates that have folivory-heavy diets have evolved gut specializations for fermentation -either through the evolution of a complex forestomach, as in colobine monkeys (Langer and Others 1988) or through the expansion of the hindgut (caecum and colon) (Cork 1996). Predictably, hindgut fermenters have caecum/colon volumes that correlate positively with the proportion of leaves that make up their total diet (Chivers and Hladik 1980). We propose that external fermentation may represent a parallel, alternative adaptation.

External fermentation supports gut fermentation
The third mechanism by which external fermentation supports digestion is by supporting and contributing to the gut microflora, which in turn contributes to ongoing enhanced nutrient absorption. It may effectively act as an external reservoir of bacteria necessary for internal fermentation. In other species, this reservoir function is supplied internally by the caecum (Palestrant et al. 2004;Swidsinski et al. 2005). Caecal size is larger in Old and New World monkeys and prosimians than in anthropoids, smaller in cercopithecoid monkeys, and reduced further in hominoids; of the great apes, humans have the most reduced caecum (Scott 1980). Humans and other apes, however, possess a vermiform appendix, located adjacent to the caecum (Scott 1980), which has been proposed to function as a reservoir for beneficial intestinal flora (Randal Bollinger et al. 2007), along with small crypts within the colon (  In summary, the ingestion of fermented foods provides four critical components to digestion and absorption. First, it increases the digestibility of foods; second, it increases the bioavailability of micronutrients; third, it supports gut fermentation by contributing to host microfloral diversity; and lastly, it supports immune function and resistance to disruption of the gut microbiome. These benefits would have been adaptive advantages for our early ancestors, and could have played a key role in human brain evolution, as we describe below.

III. External Fermentation as a Driver of Hominin Brain Expansion
The development of external fermentation technology represents a plausible metabolic mechanism leading to brain expansion beginning at our ancestors' divergence from the australopiths. The importance of considering metabolic costs in brain evolution was famously outlined in the Expensive Tissue Hypothesis, in which the reduction of gut tissue in the human lineage permits the reallocation of metabolic resources towards brain tissue, which is metabolically expensive (Aiello and Wheeler 1995). The obvious paradox here is that gut tissue, while metabolically expensive as well, is the site of caloric uptake for the organism. Thus, reduced gut sizes could only evolve if our ancestors were able to exploit a more nutrient-dense and easily-digestible food source. Following mixed results in empirical tests of the Expensive Tissue Hypothesis, Isler and Van Schaik (2014) expanded and elaborated this framework to include tradeoffs between multiple constraints and drivers in brain size evolution, of which metabolic costs and dietary changes are one component. Importantly, while there is debate and disagreement on the extent to which a gut-brain tradeoff is a causal mechanism in brain enlargement in animals generally or in humans specifically, there is widespread agreement that the evolution of larger brains is closely tied to changes in the budget of energy costs and expenditures.
Aiello and Wheeler examined the relative proportion of the most metabolically expensive tissues outside of the brain: the heart, liver, kidneys, and gastrointestinal tract. This led them to observe that the gastrointestinal tract -stomach, small, and large intestine -was 60% smaller than predicted for a primate of our size (Aiello and Wheeler 1995). But if we take a closer look at the gastrointestinal tract, we find that the reduction in size is not equal across organs. The volume of large intestine in non-human great apes is twice that of the small intestine (in gorillas, close to five times the volume); whereas in humans, the ratio is reversed, with the colon having approximately one-third the volume of the small intestine (Katharine Milton 1987; Katharine Milton 1999). Aiello and Wheeler compared the expected organ proportions for a primate of our size (using chiefly great apes for comparison) with the observed proportions and found a large difference -however, this difference was not broken down by subcomponents of the gastrointestinal tract.
Using estimations from Milton (K. Milton 1986;Katharine Milton 1987;K. Milton and Demment 1988) on differences between the proportions of small intestine and colon in humans and apes, we calculated the approximate masses of these subcomponents by taking the midpoint values given by Milton (1999) and applying them to the total gastrointestinal tract values from Aiello and Wheeler (1995). Table 2 shows these calculations; Figure 2 shows the relationship between different organ sizes in a hypothetical 65 kg human with ape-like organ sizes (expected) and the actual proportions of these organ sizes in modern western humans (actual). While total gut Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 6 October 2020 doi:10.20944/preprints202010.0135.v1 reduction is impressive (a reduction of over 41%), when we look at subcomponents, it's clear that the reduction is not consistent across the board. Small intestine proportion actually increases, from approximately .4 kg to .62 kg in modern humans, an increase of 58%. The subcomponent which accounts for the largest share of the reduction is the colon. With a predicted ape-like value of 0.85 kg, a typical human instead has an estimated mass of .22 kg, a reduction of 74% -the largest reduction of any of the gut subcomponents and any of the other major organs analyzed ( Table 2).
What permitted the drastic reduction in colon size in the human lineage? Milton has implicated meat-eating ( Milton 1999). It is reasonable to postulate that a smaller colon would reflect a reduction of dependence on fibrous plant material, given that a major function of the colon is to house bacteria that aid in the breakdown of carbohydrates to SCFAs, as described above. This idea is supported by the fact that humans and members of the order Carnivora share small colon size. However, the gut transit time in Carnivora is much faster than in humans. Milton postulates that this difference is due to our evolutionary history as plant eaters, and that meat-eating is nonetheless the most likely candidate for providing the greater nutrient density needed as gut size reduced.
An equally or perhaps more probable explanation is that colon reduction does indeed follow from reduced need to break down fibrous plant material within the digestive tract, but that this reduced need is due not to an increased dependence on meat, but rather to an increase in bioavailability of nutrients before food is consumed --i.e., external fermentation (Figure 1).

Plausibility in early hominin lifestyles
Is this scenario realistically plausible for our australopith ancestors? In our view, the major hurdle is that it requires a cache of food to be stored in a location conducive to fermentation, and remain there for a duration sufficient for it to occur. Notably, the transport and caching of food is something that separates human ancestors from our closest extant primate relatives. Early hominins appeared to have carried food resources to specific locations, as evidenced by concentrations of animal bones in larger than expected quantities; further, evidence suggests stone tools were carried large distances as well, up to 10 kilometers (Potts 1984;Toth and Schick 2009).
Combined with the accumulating evidence that stone tools were likely knapped prior to leading to the modern day, where cumulative culture has led to a remarkable diversity of fermentation practices (see Table 3). We propose that the offloading of fermentation to the extra-somatic environment can provide explanations for evolutionary changes in the human lineage: the increase in brain size as well as the reduction in gut size, particularly the large intestine. The emergence of meat-eating, tuber-harvesting, and cooking have all also been proposed to account for these changes; why should our just-so story be given any additional credence? Below, we consider several explanatory advantages of the external fermentation hypothesis versus other current hypotheses.

Less brainpower required
In searching for an initial trigger to the upward spiral of human brain expansion, it is important to recognize that it would have to occur in organisms with brains roughly the size of a chimpanzee. The cognitive capacities of chimps may arguably be inferior to those of australopiths, particularly later, larger-brained australopiths. At a minimum, though, we can reason that behaviors which are well within the chimp repertoire were likely to have been attainable by australopiths, and that behaviors which are beyond the chimp repertoire may have at least been challenging for australopiths.
Chimpanzees display a variety of complex, socially learned, instrumental behaviors oriented toward food, such as "fishing" for termites or honey using sticks, and fashioning spears to hunt monkeys. Notably, among apes, behavioral adaptations to increase net caloric gain are not limited to chimpanzees; for example, gorillas fold leaves in complex ways which increase the efficiency of consumption (Byrne and Byrne, 1993). Perhaps the most well-studied example, though, is chimpanzee nut cracking. Juvenile chimpanzees spend years learning to accomplish this using a hammer stone and anvil stone. During this time, they make errors like banging the hammer stone on the anvil stone while the nut is left resting on the ground nearby (Hirata, Morimura, and Houki 2009). This suggests that chimpanzees have difficulty understanding the underlying causal mechanismi.e., that the nut's shell is opened because it was struck. Despite nut-cracking occurring in a social context with multiple expert and novice crackers in the same location, using the same tools, at the same time, understanding of the causal relationship between percussion and a cracked shell is not socially learned. Instead, each chimpanzee independently "re-discovers" this causal relationship for itself. The social context merely contributes a scaffold in which independent learning can occur (Tennie et al., 2009).
Chimpanzee stone tool use has continued substantially unchanged for at least 4,300 years, as indicated by the discovery of fossil stone tools (Mercader et al., 2007). Thus, animals with brains similarly sized to australopiths are capable of socially transmitting instrumental behaviors which are stable over long periods of time in the absence of underlying causal understanding about how the specific details of the action are related to its end goal. Aspects of behavior that are easily socially transferred by chimpanzees include memory for the objects, tools, and locations that are involved in achieving a particular goal. We propose that this is all that is required for social transmission of fermentation to take hold.
In comparison with fermentation, the means-ends dependencies between objects, actions, and outcomes in cooking are considerably more constrained and complex. Cooking requires comprehension of causal mechanisms between multiple interacting objects -i.e., a chain of sequential, dependent interactions between fuel, flames, and raw food. This is precisely the type of means-ends dependency that is challenging for chimpanzees. Thus, we propose that external fermentation poses less of a cognitive hurdle than control of fire, and was thus more likely than cooking to impact the gut-brain tradeoff at an earlier point in evolution.
Notably, one experiment did address whether chimpanzees might have some of the cognitive skills necessary for cooking. Warenken and Rosati (2015) presented chimps with a device which, via unseen experimenter manipulation, "transformed" raw food to cooked food, and showed that chimps deliberately used the device to obtain the latter. Beran et al (2016) argue that this experiment reveals more about chimps' food preferences and capacity for bartering or exchange behavior than it does about their capacity for cooking. We propose an alternative but not necessarily mutually exclusive view, namely, that these results provide evidence that chimp-sized brains are capable of understanding and performing the steps required to ferment food: put food in a particular place, wait for it to become transformed, and then enjoy an improved version.
Why is it, then, that nonhuman apes don't engage in external fermentation? A potential explanation may be that despite having the basic cognitive mechanisms, chimps lack other, prerequisite, psychological mechanisms required for food caching to occur. One of these may be the ability to delay gratification. Other possibilities include increased social tolerance, food-sharing, and the existence of cultural norms about ownership. These mechanisms are uniquely developed in humans, which are the only ape species known to store food, and may be necessary for the continued existence of a cache of communally accessible food.

No lightbulb moment required
While the utility of fire and fermentation for food processing could both be discovered accidentally, we argue that this discovery was more probable for fermentation. Naturally-occurring fire is not a daily incident. Opportunities for our ancestors to spontaneously observe fire or notice its potential for cooking must have been sporadic. It is conceivable that accidental cooking may have occurred (for example, the action of wildfire on animal carcasses or buried tubers), but this seems likely to have been infrequent. More importantly, the transition from opportunistic, infrequent access to accidentally-cooked food to a long-term and stable source of extra calories would require a "lightbulb moment:" recognition of the effects of the accidental process, and intentional, deliberate actions to reproduce their causes. In contrast, naturally-occurring fermentation is a daily incident. Bacteria and fungi are everywhere, all the time, and spontaneously colonize food that isn't consumed or otherwise preserved. Moreover, no "lightbulb moment" is required to transform unintentional external fermentation into a stable, ongoing source of extra calories.

Environmental stability
Fires require ongoing active effort to maintain, whereas fermentation is largely a passive process. Once started, an ongoing fermentation does not extinguish, and does not require tending or restarting, as fire does. Moreover, this environmental persistence offers more chances for social learning, further supporting the longevity of the practice across generations.

Stable food preservation -a caloric buffer
Because brain tissue is so energetically expensive, and is intolerant of reduced energy availability, organisms with larger brains are more susceptible to fluctuating availability of food (van Schaik). The evolution of increased adipose tissue in humans is a proposed adaptation to ameliorate this risk; fat provides an "internal buffer" for survival through lean times (Leonard et al., 2003;Navarette et al., 2011). External fermentation practices may have provided a secondary, "external buffer." Fermentation can preserve food for years. Food spoilage is caused by microorganisms, and some of the best inhibitors of microorganisms are other microorganisms. Fermentation allows for the proliferation of non-harmful or beneficial strains which out-compete harmful strains; for example, by-products of fermentation include alcohol and acid, which inhibit further microbial growth, effectively preserving the food. There are other food storage techniques whose effective timescales are within that of fermentation, such as smoking, drying, freezing, and salting (notably, often used in combination with fermentation).
However, compared to these other methods, we propose that fermentation may have been accomplishable more easily, across a wider range of environments, and by earlier, smaller-brained, less cognitively-complex ancestors. Fermentation accounts for all the benefits that cooked food offers: softer food, higher caloric content, greater bioavailability of nutrients, and protection from pathogenic microorganisms. Fermentation solves several problems. It does not require special materials beyond a place to store food (a hollow, a cave, or a hole in the ground work).

Summary of explanatory power of the External Fermentation Hypothesis
It does not require overcoming fear -there is a low barrier to entry. It can be stumbled upon rather than requiring planning and tool use. And it does not require, initially, longterm planning, focused attention, or sophisticated social coordination.
In all likelihood, for most of human history, it was nearly impossible to store food for any length of time without bacterial or fungal growth. Life-threatening illness is a risk of some food-borne microbes (e.g., E. coli, salmonella). Thus, it would have been necessary to either keep all microbial growth below potentially toxic levels (via e.g., drying, salting, smoking, or freezing), or encourage high levels of "good" microbial activity to out-compete the bad. The latter seems clearly easier.

Contemporary Human Fermentation Practices
We can look to current fermentation practices for insight into its role in our past. We have created a detailed list of examples that provide a sense of the widespread scope and impact of fermentation technology on the human diet worldwide ( Table 3). Humans deliberately ferment foods of nearly every kind, including fruits, vegetables, grains, legumes, animals (muscle meat, organs, fat and bones), dairy, fish, and shellfish.
Fermentation is practiced successfully in a diversity of climatic contexts, from tropical humid conditions to arctic environments. It is accomplished with a wide variety of microorganisms, including bacteria, filamentous fungi, and yeasts. Moreover, fermentation works on a range of timescales from hours to years; it can effectively act as a short-term flavor enhancer or a long-term storage technique.
We present this aggregation of examples as evidence supporting three points. First, given the incredible range of food types and environments that can lead to successful fermentation, it is plausible that this was also possible for the food types and environments of early human ancestors. Second, it seems that fermentation is ubiquitous across extant cultures and can be considered a human universal. This is consistent with fermentation having a very early emergence. Third, while cultural practices for fermenting food vary across the globe, it seems clear that humans in general have a taste for fermented food. This preference may be an evolved mechanism which emerged because an attraction to these flavors was adaptive in our shared past. Notably, many fermented foods listed in Table 3 such as fish sauce, soy sauce, and vinegar, are condiments -i.e., substances added to other food items mainly for the purpose of improving palatability.

V. Testing the External Fermentation Hypothesis
Are preferences for fermented foods innate?
If our hypothesis is correct, then we might expect to find evolved innate preferences for beneficial fermentation products, or evolved innate aversions to dangerous byproducts of "off" fermentation. Interestingly, it appears that many of the most disparately-regarded foods -seen by some as prized delicacies, and by others as supremely unappetizingare fermented: for example, thousand-year eggs, natto, and Limburger cheese. These preferences appear to be highly culturally-specific, which might be adaptive given the high cultural diversity of fermentation practices and the risks of consuming a ferment gone awry. The same flavors or odors which might signal "good" food in one culture could emanate from "off" ferments in another. Future research could address the extent to which preferences for fermented products are innate, cultural, or may be the product of gene-culture coevolution (Henrich and McElreath 2003). Are they more susceptible to cultural learning than other food preferences? Are they more sensitive to experience in a developmental critical period, and/or less flexible after this period closes? Are they heritable, either genetically or epigenetically (Dias et al. 2015)?

Do the risks of external fermentation outweigh the benefits?
A potential argument against our hypothesis concerns the potential for fermented foods to be colonized by pathogenic microbes. This must certainly have been a risk, but the more relevant question is, how did the risks and benefits of external fermentation compare to the risks and benefits of other potential solutions to the of balancing the metabolic budgetary increase associated with brain enlargement? Hunting, scavenging from large carnivores, and use of fire certainly carry their own risks; perhaps the risks of fermentation were more predictable and thus more reliably mitigable through individual and cultural learning. In the environments and time periods relevant for our hypothesis, what kind of situations might have caused a fermentation to go "off"? How easy would it have been for a hominid with a chimpanzee-sized brain to avoid these risks, either deliberately or via socially-learned canalization of practices? How often would "off"

Is the human microbiome adapted for consuming fermented food?
Another opportunity to find evidence for or against the External Fermentation Hypothesis may come from examinations of the human microbiome. Interestingly, a comparative analysis with chimpanzees, bonobos, and gorillas indicates that the human microbiome has undergone accelerated deviation from the ancestral ape state, and now shows reduced diversity (Moeller et al., 2014), which may be consistent with increased reliance on external microbial communities. If early humans really offloaded internal fermentation to the external environment, we should expect to see changes in the internal microbial community, and potentially evolutionary cross-talk between internal and external fermenters of human food. Would internal species associated with a particular food become less abundant over time, while the external species proliferated?
Would humans' internal flora adapt to now specialize in the post-fermentation product, perhaps with evolved adaptations for tolerating higher levels of fermentation byproducts like acid or ethanol? Can we trace the co-evolution of gut flora and external fermentation flora as human populations have moved around the globe? Could phylogenetic analyses of human gut microbes provide a window onto the onset of fermentation practices in human evolution?

Conclusions
We have proposed that the acquisition of fermentation technology by early homininsthe external fermentation hypothesis -is a good candidate mechanism for human brain expansion and gut reduction. The offloading of gut fermentation into an external cultural practice may have been an important hominin innovation that laid out the metabolic conditions necessary for selection for brain expansion to take hold. While the potential importance of fermentation in the evolving human diet has recently been postulated (Dunn et al., 2020), and the reduction in human colon size has long been known (Milton, 1999), to our knowledge, the possibility that external fermentation served as the initial trigger in the human lineage for the expansion of brains and the reduction of the gutspecifically, the colonhas so far been unnoticed. We have discussed the adaptive benefits of this hypothesized scenario, its realistic plausibility, and its explanatory power relative to other hypotheses. We invite commentary and experimental tests from the broader academic community.  (McNeil 1984;Livesey 1995;McBurney 1994) Legend: Percentage of maintenance energy derived from the production of short-chained fatty acids (SCFAs) via gut fermentation. Information adapted chiefly from (Bergman 1990). Values for gorillas were estimated from diet composition and human colonic fermentation rates. Data based on Aiello and Wheeler's (1995) compilation of data from Stahl (Stahl 1965), Stephan et al. (Stephan, Frahm, and Baron 1981), and Chivers and Hadlick (Chivers and Hladik 1980). Gastrointestinal tract weights were subdivided based on ratios from Milton (K. Milton 1986;Katharine Milton 1987; K. Milton and Demment 1988).     Table 1. "Expected" represents the ratio of organ masses expected if humans had proportions in line with other great apes. "Actual" represents an estimation of the ratios in a typical modern Western human.