In SeungkiKwak/Kwak_S_PhD_thesis: Sungki Kwak's PhD Thesis

set_parent("thesis.Rmd")

\chapter{Discussion}

\section{Introduction}

In this chapter I will further explore the initial interpretations that I made in chapter five. First, I will re-evaluate the current rice-based model and its assumptions about the strict dichotomy between Chulmun hunter-gatherers and Mumun full-dress rice farmers. Then, by correlating the results of the organic geochemical analyses and luminescence dating with available bulk isotope and paleobotanical data, I will make an argument that the role of rice as a subsistence strategy in the central part of the Korean Peninsula was relatively more minor than previously argued. Lastly, I will discuss the important implications of the results in this thesis.

\section{The transition from Chulmun to Mumun revisited}

Before I move on to the discussion about the results of the organic geochemical analyses and luminescence dating from the four sites, now is a good moment to revisit the prevailing concepts of the Chulmun and Mumun periods in the Korean archaeology, focusing especially on potteries. As I indicated in chapter two, the potteries of the two periods have several key physical traits which have lead Korean archaeologists to consider the discrepancy between those of one period and those of the other. For example, the fundamental characteristics of the Chulmun period potteries are the comb-shape pattern and pointed bottom (Figure \ref{CMpottery}a). On the other hand, all the Mumun period potteries have the flat bottom, and the major part of their body does not have any pattern. In some cases some patterns still exist, but are confined to the extreme upper part of the body.

At a first glance it sounds quite reasonable to divide the two periods, especially when we compare the Chulmun potteries showing the extensive and intensive comb-shape pattern, and the mostly un-patterned Mumun potteries (Figure \ref{CMpottery}). However, if we examine closer, there are some variations in characteristics which have been somewhat neglected. Until now, probably the most well-known Chulmun period pottery is the one from Amsa-Dong (Figure \ref{site_locations}; \ref{CMpottery}a). The entire body of a pot is decorated with comb-shape patterns which can be divided into three different parts (Figure \ref{Cpottery}a). The pattern in each part has a different length and a different angle which makes a distinctive characteristic of the pottery. This was the earliest form of the Chulmun potteries in the central part of the Korean Peninsula, and appeared around 6000 BP. However, the pattern on the Chulmun potteries gradually changes as time goes by. Figure \ref{Cpottery}b presents a Chulmun period comb-shape patterned pot found at an upper layer of the Amsa-Dong site. Interestingly enough, the patterns on its bottom part are gone and those of its middle part became less distinctive. If we see the Late Chulmun period potteries excavated from the Amsa-Dong site, Seoul city and the Sammok island site, Incheon city (Figure \ref{Cpottery}c; \ref{Cpottery}d), we can verify the patterns on their middle part also vanished away and only those on their top part remain. These latest Chulmun potteries from Sammok island and Amsa-Dong give an important clue to the relationship between the Chulmun and Mumun potteries: the pattern in this stage of the Chulmun potteries exists only on their top and rim, and the most distinctive characteristic of the Incipient/Early Mumun potteries is various patterns on their rim (Figure \ref{Mpatterns}c). These rim-based decorations make the two potteries seem similar. Some of the Chulmun potteries have even the rim-punctuation (Figure \ref{Cpottery}e) which is commonly observed on the Early Mumun potteries (Figure \ref{Mpatterns}c). These similarities in pattern between the late Chulmun, and the Incipient/Early Mumun potteries suggest a close connection between them. All this contrasts with the current dominant idea which assumes the discrepancy between the Chulmun and Mumun periods [@Kim2006a; @Kim2003; -@Kim2006].

\section{The Chulmun and Mumun periods: Essentialism vs. Materialism}

If we consider the relationship between the Chulmun and Mumun potteries in light of the two different perspectives of essentialism and materialism, the picture is clearer. Essentialism is a philosophical stance which supposes the existence of a specific entity which can be both identifiable and distinguishable. If things share actual and fixed characteristics, these essential traits can be used to distinguish group A from another group B, and constitute the essence of each of the two groups. The most prominent point of this ‘essence’ is its unchanging permanency. In contrast to essentialism, materialism holds that phenomena cannot exist as fixed entities, because they are always in the process of becoming something else. In materialists’ view, things are in a state of flow: no two things can ever be put into the same category, because even similar things are just at similar points in the process of becoming others. In the discipline of biology, the two philosophical stances were identified by Ernst Mayr [-@Mayr1959] in defining the concepts of biological species: ‘typological’ thinking versus ‘population’ thinking [@Marwick2008, Figure \ref{MvsE}]. A key point in differentiating essentialism from materialism is not that the former treats difference and the latter change, but that the one treats only difference while the other treats both difference and change [@OBrien1998: p. 29].

For many years, archaeologists in Korea have been studying sites and artifacts within the framework of ontological essentialism by the name of ‘typology’ or ‘classification’. Archaeologists often use the terms classification and typology interchangeably, but a distinction must be made between them. A classification is any set of formal categories into which a particular field of data is partitioned, while a typology is a particular type of rigorous classification, in which a field of data is divided up into the categories that are all defined according to the same set of criteria, and that are mutually exclusive [@Adams2001; @Adams2007]. Therefore, to be precise, it is not classification but typology that many archaeologists have been employing for studying their sites and artifacts. They grouped artifacts according to some characteristics for demonstration of cultural traits and cultural changes. Objects are split into categories —in other word, ‘types’ — according to their perceived similarities, and change is viewed as transition from one type to another. This means that as long as the objects are in the same category, they are closer to each other than any other objects in different categories (even if one category comes right after another in terms of time). By doing so, archaeologists have been creating units for their interpretation of archaeological phenomena. However, there is nothing “inherent” in the units they use that makes them real [@OBrien1998: p. 30]. Artifacts may share certain traits in common which make us put them in the same category, but there is no reason to think each category is genuine.

A similar trend has been prevailing in the Korean archaeology, especially when we consider the transition from Chulmun to Mumun. Whether they recognized or not, Korean archaeologists created two units, named them “Chulmun pottery” and “Mumun pottery”, and regarded the two as different entities. In addition to that, they expanded this concept to the entire archaeological phenomena of the two periods. Since change is viewed as transition from one separate entity (Chulmun) to another (Mumun), everything in the Chulmun period has to be drastically different from everything of the Mumun period, including pottery, house pits, stone artifacts and, of course, subsistence strategies [cf. @Kitcher1981; -@Kitcher1989; @Wylie2002; @Strevens2004].

I do admit that I may have somewhat exaggerated about how Korean archaeologists understand the relationship between Chulmun and Mumun. Beside, we have to consider the possibility of sudden and rapid transition from the former to the latter which caused the actual disconnection between the two at least in certain areas by human migrations [@Kim2003; -@Kim2006b]. Anyhow, we have observed the connection between Chulmun and Mumun through the examination of changes in pattern on the potteries excavated from the central part of the Korean Peninsula (cf. Figure \ref{Cpottery}). In fact, there are archaeologists who have already recognized the similarities between the patterns on the potteries from the two periods [@Shin2007]. Nevertheless, many archaeologists in Korea still excavate and investigate their sites focusing on the differences between the overall archaeological assemblages of the two periods, and mostly adopt the migration model [@Kim2003; -@Kim2006b] to justify the artificial units over the real variation.

Frankly speaking, creating units for the interpretation of archaeological phenomena is somewhat necessary. Just like the scales which show mass or length — kilograms, meters, inches — archaeological units are useful for analysts for documenting the variation across the real things [@OBrien1998]. However, we should be aware that archaeological records must be understood as a ‘continuous’ sedimentary process, and concentrations of artifacts are the products of numerous events of deposition. Only with this awareness, units (e.g. Chulmun and Mumun) can be used as means of measurement.

\section{The subsistence of the Chulmun and Mumun periods}

We have observed the connection between Chulmun and Mumun through the examination of changes in pattern on the potteries (cf. Figure \ref{Cpottery}). What, then, about the subsistence of the Chulmun and Mumun peoples? Were the foodstuffs of the Chulmun people quite different from those of the Mumun people? What was the role of rice? Did the Mumun people really rely heavily on rice as previously suggested? Was rice a hallmark of the Mumun period?

\subsection{The Chulmun subsistence}

Traditionally, what we know about the subsistence of the Chulmun people is that they relied mainly on hunting and gathering; but from the Middle stage (5,500 BP) of the Chulmun period, we are able to observe evidence of the initial domestication of foxtail and broom-corn millet [@Norton2007]. Recently. G. Lee [-@Lee2011b: p. S326] argued that they had specific subsistence solutions which include combinations of wild (e.g. acorn (QuercusacutissimaCarr.), Manchurian walnut (Juglans spp.)), possibly managed (e.g. chenopod (Chenopodium sp.), panicoid grass (Paniceae)), and domesticated (e.g. foxtail (Setariaitalica ssp. italica) and broomcorn millet (Panicummiliaceum), possibly soybean (Glycine max), azuki (Vignaaugularis) and beefsteak plant (Perillafrutescens (L.) Britt)) plants. However, before we define the subsistence of the Chulmun period, it is worth to examine available bulk isotope data.

Throughout the Chulmun period, ancient people occupied the coastline of almost the entire Korean Peninsula, and created shell middens in many different locations near coastal areas. Because of these shell middens, it is easier for us to trace the ancient diet, for they provide excellent environment in terms of preservation of organic materials. Numbers of bulk Isotope studies have been conducted, since both human and animal bones were excavated from these middens [@Ahn2006; @Choy2010; @Kim2010].

The isotope analysis on the human remains and animal bones excavated from the Geoje and Tongsam-Dong shell middens, southern part of the Korean Peninsula (Figure \ref{site_locations}) shows a focused diet of the Chulmun people and terrestrial animals [@Kim2010; @Choy2010]. The relatively low $\delta$^13^C and $\delta$^15^N values of wild terrestrial mammals such as pigs and deer indicate that the diets of these animals were dominated by C~3~ plants (Figure \ref{bulk1}), which means most of indigenous wild plants in the Korean Peninsula are C~3~ plants [cf. @Lee2011a]. On the other hand, the $\delta$^13^C and $\delta$^15^N values of the human bone collagen were quite close to those of marine animals (Figure \ref{bulk1}), indicating the Chulmun people mainly consumed marine resources. It suggests, along with the geographic location of the middens (Figure \ref{site_locations}), that procuring marine resources was their main subsistence strategy, though terrestrial mammals were included in their diet.

\begin{figure} \centering \includegraphics{figures/bulk1.jpg} \caption{The results of the bulk isotope analysis on human remains and animal bones excavated from the Daepo and Tongsam-Dong shell middens (modified from J.J. Lee 2011b: p. 41)} \label{bulk1} \end{figure}

D. Ahn [-@Ahn2006, Figure \ref{bulk2}] conducted isotope analysis on the human remains and animal bones for the Konam-Ri shell midden (Figure \ref{study-area_Konam-ri}b), central part of the Korean Peninsula. The samples were collected from both the Chulmun and Mumun periods. The isotope analysis on pig bones from both periods indicates that the pigs mainly consumed C~3~ plants. The $\delta$^13^C and $\delta$^15^N values of the human bone collagen from the Chulmun period were close to those of a wild pig, which means the Chulmun people mainly relied on wild C~3~ plants as well as wild pigs. However, considering the difference in $\delta$^13^C values between human (-17.7 $\permil$) and pig (-19.7 to -21.2 $\permil$), we cannot totally eliminate the possibility of C~4~ plants in the human diet.

Based on the results of isotope analyses and paleobotanical studies, one may conclude that the Chulmun people mainly relied on hunted terrestrial animals/marine mammals, gathered wild C~3~ plants and, in some degree, domesticated C~4~ plants.

\subsection{The Mumun subsistence}

One of the most debated issues related to the Mumun subsistence in the central part of the Korean Peninsula is the role of the intensive agriculture heavily based on rice. On one hand, the intensive agriculture was viewed as "cure-all remedy" [@Lee2011b: p. S327] which substituted for any other subsistence resources [@Lee2001; @Norton2000; @Kim2003; -@Kim2006b]. In addition to that, the emergence of a social hierarchy and the subsequent social complexity were considered to be driven by the rapid spread of the intensive rice agriculture into foraging contexts [@Kim2006a; -@Kim2006]. On the other hand, the role of both broom corn and foxtail millet was emphasized together with that of azuki and soybean [@Crawford2003; @Lee2011b]. In the latter view represented by paleobotanical studies, contrary to the former, rice agriculture was regarded as no more than an 'add on' to the existing millet-based subsistence originating from the Chulmun period. However, the results described in the previous chapter of the organic geochemical analyses of potsherds from major Mumun habitation sites told a story somewhat different from either of these existing arguments.

Table \ref{tab:Mumun_c3_animal} shows the proportions of the potsherds indicating the presence of C~3~ plant oil to all the analyzed samples from the three Mumun sites studied in this thesis. More specifically, the table shows the percentages of the samples interpreted as containing traces of C~3~ plant oil to the total number of the analyzed samples from each site. As time goes by from the Early Mumun to the Middle/late Mumun period, the proportion of C~3~ plant oil increases from 14% to 38% (cf. Figure \ref{Radiocarbon_dates}). However, this does not mean that the proportion of rice in the Mumun farmers' diet increased from 14% to 38%. As I mentioned in Chapter five, C~3~ plants include not only rice, but also legumes and barley. We have the pollen data from 5,500 BP to 2,600 BP showing that the ancient farmers of the Korean Peninsula utilized soybean (Glycine max) and azuki (Vignaaugularis) as subsistence resources [@Lee2011b]. Therefore, it is impetuous to argue that the detected C~3~ plant oil is from rice alone. In addition to that, most of the samples identified as revealing C~3~ plant oil are also interpreted as containing pork adipose, ruminant adipose, or freshwater resources. This is because the ellipses of C~3~ plant oil, pork adipose and ruminant adipose overlap one another (cf. Figure \ref{CSIAKimpo}; \ref{CSIAKM2}; \ref{overlap}), and if the $\delta$^13^C values are plotted in the area where the reference ellipses for CSIA overlap, we cannot pinpoint a single food class as origin of the $\delta$^13^C values.

Even if we regard the origin of all these $\delta$^13^C values from the 16 samples (cf. Table \ref{tab:Mumun_c3_animal}) as C~3~ plant oil, and also suppose all C~3~ plant oil is from rice alone, the proportion of rice in the entire diet is below 40% at most. The overall proportion of the rice revealed by the entire analyzed samples in this thesis, which is relatively low, strongly suggests that rice agriculture might not have been a major subsistence strategy throughout the Mumun period. However, we need to consider with what percentage, we can say rice agriculture was the dominant subsistence strategy. Someone might say 38% is already enough for arguing rice as the major subsistence resource. But, before making the final decision, it is good to look into the archaeological context. The site that showed 38% of rice among all food classes is Songguk-Ri (Middle Mumun period). Archaeologists have been arguing that rice became the hallmark of the ancient farmers’ subsistence by the Songguk-Ri stage, and acted as a trigger of emergence of a social hierarchy and the subsequent social complexity [@Kim2006a; -@Kim2006]. Yes, 38% is not a small proportion and I do think rice might have been one of the main subsistence subsistence especially from around 2600 BP. Nevertheless, considering these prevailing thoughts of the role of rice during the Middle Mumun period, I think the results from Songguk-Ri do not fully support the idea of rice as dominant subsistence strategy.

At this point, I have to admit that the quantitative proportion of different food classes, which is derived from the analyses, cannot be assumed to be the direct representation of the diet taken by the ancient farmers. For example, even if the ratio of C~3~ plants to porcine fat of a site is 1:3, this does not mean its ancient farmers relied exactly 3 times more on pork than rice. Therefore, it is reasonable that the results of the analyses in this thesis are viewed as a macroscopic explanation.

\begin{table}[h] \begin{tabular}{p{6cm}p{3cm}p{3cm}p{3cm}} \hline & Sosa-Dong & Kimpo-Yangchon & Songguk-Ri \ \hline Main occupational period & 3000-2700 BP & 3000-2700 BP & 2500-2300 BP \ Proportion of C~3~ plants oil (Max.) & 14 \% (4 samples) & 25 \% (5 samples) & 38 \% (7 samples) \ Proportion of terrestrial animals and aquatic resources (Min.) & 71 \% (20 samples) & 60 \% (12 samples) & 61 \% (11 samples) \ \hline \end{tabular} \caption{The proportion of C3 plant oil and terrestrial animals/aquatic resources from the three Mumun sites studied in this thesis (percentage of the number of the samples interpreted as containing traces of C3 plant and terrestrial animals/aquatic resources to the total number of the analyzed samples from each site)} \label{tab:Mumun_c3_animal} \end{table}

If rice (C~3~ plants) agriculture was not the major subsistence strategy of the Mumun people, what about millet (C~4~ plants)? Based on archaeobotanical analyses on both Chulmun and Mumun sites, G. Lee [@Lee2011b; @Crawford2003] stressed the role of both broom corn and foxtail millet as major subsistence resources of the Mumun farmers. Since the study of G. Lee was based on well-organized systematic paleobotanical analyses, the presence of C~4~ plant oil was highly anticipated in my analyses. Surprisingly, among the 113 samples collected from three Mumun sites, only one sample showed $\delta$^13^C values close to the C~4~ range (KIM049: palmitic acid: -16.8 $\permil$, stearic acid: -17.2 $\permil$). However, it is yet to be confirmed whether this one sample indicates the presence of millet. Recent studies in China revealed that the $\delta$^13^C range of modern millet is from -10.48 to -10.05 $\permil$, higher than the average range of C~4~ plants (-17 to -12.5 $\permil$) [@Pechenkina2005; cf. @Malainey2010].

There could be several explanations about the absence of C~4~ plant oil. Firstly, it is possible that millet was processed/cooked without water; for example, popped or roasted [cf. @Reber2004]. Especially, considering the size of the millet grain which is smaller than that of the rice grain, popping might have been the major cooking method for millet, for it enlarges the size of the grain. Another possible explanation is that millet was often cooked with other food stuffs. One may assume that millet grains may have been cooked together with pork or marine animals as an additive ingredient because of their small size. If millet was cooked in a pot along with other resources, C16:0 and C18:0 fatty acids from this pot may indicate relatively low $\delta$^13^C values in comparison with those of millet. The result of the bulk isotope analysis on human remains from the Mumun period ($\delta$^13^C: –12.2 $\permil$, $\delta$^15^N: 10.1 $\permil$) might indicate this type of cooking method [@Ahn2006, Figure \ref{bulk2}]. Relatively low $\delta$^13^C value in comparison with those of modern millet (from -10.48 to -10.05 $\permil$) and somewhat high $\delta$^15^N value indicate combination of millet, terrestrial mammal and marine resources.

However, it is also possible that the absence of C~4~ plant oil simply means millet was not the main subsistence resource throughout the Mumun period. Though the study of D. Ahn [-@Ahn2006] showed the evidence of a C~4~ plant in the Mumun people’s diet, the data that it utilized was based only on one human remains from the Konam-Ri shell midden (Figure \ref{bulk2}), and so its result does not have pertinence in terms of representing the overall Mumun subsistence. Indeed, a limited number of cases cannot represent the dietary pattern of the entire Mumun population.

If both rice (C~3~ plant) and millet (C~4~ plant) were not main foodstuffs throughout the Mumun period, what was the most wide spread and reliable subsistence strategy of these ancient farmers? Table \ref{tab:Mumun_c3_animal} shows the proportions of the potsherds indicating the presence of terrestrial animals and aquatic resources (marine/freshwater) to all the analyzed samples from the three Mumun sites. Note that these data exclude all the $\delta$^13^C values that might have originated from C~3~ plant oil (rice). As I mentioned above, the reference ellipses of C~3~ plant oil, pork adipose and ruminant adipose overlap one another (cf. Figure \ref{CSIAKimpo}; \ref{CSIAKM2}; \ref{overlap}). Therefore, by eliminating all the $\delta$^13^C values that plotted in the area where the reference ellipses of pork and ruminant adipose overlap the ellipse of C~3~ plant oil, I can exclude all the $\delta$^13^C values that might have originated from C~3~ plant oil (e.g. rice). In this setting, the percentages of terrestrial animals and aquatic resources in Table \ref{tab:Mumun_c3_animal} can be considered as the most conservative estimates. Even from the most conservative viewpoint, the diets of all the three Mumun sites were dominated by terrestrial animals and aquatic resources. The picture seems to be that hunting and fishing still persisted even a thousand years after the initial introduction of rice farming.

Especially, among the terrestrial animals pork was the major foodstuff (Figure \ref{CSIAKM2}; \ref{CSIASosa2}; \ref{CSIASongguk2}) throughout the Mumun period. However, this is not a surprising result, for we already have solid evidence of pork consumption since the Chulmun period [cf. @Lee2011; @Lee2011a]. In a recent study, J. Lee [-@Lee2011] even mentioned the symbolic significance of pork consumption in the Korean Peninsula. Whether or not there are symbolic meanings in pork consumption in these samples, it is unquestionable that pig hunting was one of the major subsistence strategies since the Chulmun period.

Traditionally, the Chulmun – Mumun transition was explained as rapid, mainly due to climate-driven human migrations from the northeastern region of China [cf. @Kim2003; -@Kim2006b]. Focusing on the overall differences in archaeological assemblages of the two periods, Korean archaeologists emphasized the discrepancy between the Chulmun and Mumun traditions. The Mumun migrants seem to be portrayed as a highly able group who could have eradicated the Chulmun indigenous foragers. Since the Mumun migrants were armed with new technology and an innovative subsistence strategy — intensive (rice) farming — they were able to spread suddenly and swiftly into the foraging contexts to leave little evidence of a transitional period [@Kim2006a]; and they were subsequently able to constrain the mobility of the indigenous hunter-gatherers by blocking their ways to resource patches to enhance the transition to farming [@Kim2003; -@Kim2006b]. Rice eventually became the hallmark of the ancient farmers’ subsistence by the middle Mumun period, and acted as a trigger of the emergence of a social hierarchy and the subsequent social complexity [@Kim2006a; -@Kim2006]. After this somewhat expedient explanation was established, archaeologists tend to focused on finding farming tools and carbonized grains (especially, rice grains).

I do admit that the results from the Middle Mumun period in this study indicated possibility of rice as one of the main subsistence strategies. Nevertheless, an overly simplified stress on rice farming blurs the real complexity of the Mumun subsistence [@Lee2011b]. In many studies, the role of the other subsistence strategies in the Mumun farmers' diet, especially hunting terrestrial animals, have largely been neglected. However, the results of the organic geochemical analyses in this thesis showed that hunting and fishing persisted still well after rice farming was introduced.

\begin{landscape} \begin{figure} \centering \includegraphics{figures/Radiocarbon_dates.jpg} \caption{Density distributions of all radiocarbon dates from each site studied in this thesis, using the R package BChron (Sosa-Dong: SS, Kimpo-Yangchon: KM, Songguk-Ri: SG, Eupha-Ri: EP) All dates were calibrated using 'intcal13' calibration curve} \label{Radiocarbon_dates} \end{figure} \end{landscape}

\section{The subsistence of the Iron Age}

In the Korean archaeology, the Iron Age is an area somewhat less studied using typical methods of archaeological science . This is partially because for the times since this period, vast documentary records from the Han Dynasty of China have been extensively employed for the interpretation of Korean archaeological phenomena. Though these documents are valuable in terms of treating the contemporary past, they are neither chronicles nor meticulous ethnographies. They offer tantalizing snippets of information, allowing variable interpretations, as can be seen in the varying discussions about this period by Chinese and Korean historians [@Nelson1993]. Under these circumstances, not much information related to the subsistence of the Iron Age was released until the systematic paleobotanical investigation and isotope analysis regarding this period began to be effected [@Choy2010; @Lee2010; @Jeong2010].

According the paleobotanical evidence given by various places in the central part of the Korean Peninsula [@Lee2010; @Jeong2010], the Iron Age people had a diet focused on C~4~ plants. They mainly consumed different kinds of millets (foxtail, broomcorn and Japanese millet). On the other hand, the isotopic evidence from the human bone collagen showed somewhat different possibilities. In 2009, Choy and Richards [-@Choy2009] conducted the bulk carbon and nitrogen analysis on 48 human bones and 45 animal bones which were excavated from the Iron Age (ca. 200 BC – 100 AD) shell midden of Nuk-Do island, Sacheon city [cf. @Seo2004, Figure \ref{site_locations}]. The $\delta$^13^C and $\delta$^15^N values indicated that the people consumed C~3~ plants, terrestrial animals and possibly marine resources. Indeed, a direct comparison between the results of the investigations conducted in the inland and an island is not recommended, for the overall subsistence strategy might be quite different between the peoples of the two regions. Anyhow, we must remark that the paleobotanical evidence cannot provide any information related to animal consumption: a further investigation has to be conducted to see if terrestrial animals or aquatic resources were also regularly consumed at inland villages of the Iron Age.

According to the results of the organic geochemical analysis of the potsherds from the Eupha-Ri site (Table 5.17; Figure \ref{CSIAEUP}; \ref{CSIAEUP2}; \ref{CSIAEUP3}), over 50% of them came from the pots that were used for processing terrestrial animals and aquatic resources. It strongly suggests the possibility that the Iron Age people regularly consumed animals such as ruminants, pork and freshwater fishes. About 25% of the analyzed potsherds showed the possibility of C~3~ plant consumption. Interestingly, no sample showed the presence of C~4~ plant. Taking into account the study of Choy and Richards [-@Choy2009], it is possible that C~4~ plants were not the main part of the Iron Age diet. However, it is also possible that potsherds from the pots used to cook C~4~ plants such as millet were simply not sampled, for the samples were collected only from eight of the total 36 excavated house pits under the limited condition described in chapter five.

\section{Luminescence dating}

As I mentioned in chapter three, the luminescence dating is more effective than the other dating methods, especially in terms of pottery chronology. One of the two main goals of this thesis is to establish a long term chronology of subsistence from the Incipient/Early Mumun period to the Iron Age (3,400 - 2,000 BP). The radiocarbon dating does not date potteries themselves but the nearby organic remains (e.g. charcoal). This means the dating event inevitably has a variable relation to the target event of pottery manufacture. The luminescence dating reveals when the pot in question was made. In order to grasp the chronology of subsistence, archaeologists need to know the details of the cooking events. Since the cooking event is more likely to be associated with the manufacturing event than the depositional event, the luminescence dating is probably the most suitable method for establishing a subsistence chronology.

I have to admit that the number of samples for the luminescence dating in this thesis is quite small to build a general chronology of subsistence from 3,400 to 2,000 BP. At the same time, I also have to admit that the sites included in this study are relatively well dated with AMS radiocarbon dating method. Since the charcoal samples for the radiocarbon dating were collected from the hearths inside of the house pits (cf. Figure \ref{CMhouse}), one may argue that the dates are relatively well associated with the cooking episodes. I do not disagree with this assumption and recognize the credibility of the published radiocarbon dates. However, on the other hand, I think we can still be benefited by the luminescence dating, due to its inherent nature of dating the manufacturing event.

Table \ref{tab:date_compare} shows the comparison between the radiocarbon dates and the luminescence dates of the four sites studied in this thesis. Note that all the dates were accumulated, including the error terms.

\begin{table}[h] \begin{tabular}{p{3cm}p{6cm}p{6cm}} \hline & Luminescence dates (accumulative; calendar year) & Radiocarbon dates (accumulative; calendar year) \ \hline Kimpo-Yangchon & 920 – 560 BC (2 dates) & 1420 – 415 BC (43 dates) \ Sosa-Dong & 790 – 280 BC (2 dates) & 1300 – 260 BC (20 dates) \ Songguk-Ri & N/A & 835 – 200 BC (18 dates) \
Eupha-Ri & 370 BC – 590 AD (3 dates) & 168 – 356 AD (4 dates) \ \hline \end{tabular} \caption{The comparison between the luminescence dates and AMS radiocarbon dates of the four sites } \label{tab:date_compare} \end{table}

The biggest challenge in here is the relatively wide error terms in the luminescence dates (Table \ref{tab:LM_KM}; \ref{tab:LM_SS}; \ref{tab:LM_EP}), in comparison with those in the radiocarbon dates (Table \ref{tab:radiocarbon_KM}; \ref{tab:radiocarbon_SS}; \ref{tab:radiocarbon_SG}; \ref{tab:radiocarbon_EP}). The average error term of the luminescence dates was about 120 years; and it created a range of error of about 240 years in each of the dates. Despite the relatively wider error terms of the luminescence dating compared to the Radiocarbon method, all the luminescence dates from the Mumun period were within the range of the Radiocarbon dates. The results from the Iron Age Eupha-Ri site was a bit different, indicating both the upper and lower limits of the luminescence dates exceeded those of the radiocarbon ones regardless of the error terms. This might be simply because the number of radiocarbon dates from the Eupha-Ri site is much smaller than those from the Mumun period sites, creating a narrower range of age.

\begin{figure} \centering \includegraphics{figures/Radiocarbon_and_Luminescence.jpg} \caption{The comparison between the AMS radiocarbon dates and luminescence dates of the four sites} \label{date_compare2} \end{figure}

Despite the small sample size and the issue mentioned above, the luminescence and radiocarbon dates from the four sites are somewhat correlated with each other overall (Figure \ref{date_compare2}). Through the luminescence dating, at least in the macroscopic view, I was able to build a general chronology of subsistence between the sites.

\section{Implications and future directions}

Since the methods employed in this thesis are based on several assumptions, there are a number of potential sources of error in this study. Future directions will focus on overcoming those potential limitations.

First, boiling food inside of a pot is not the only option for cooking. For example, as I mentioned above, it is possible that millet was processed/cooked without water; for example, popped or roasted. Though ethnographic studies showed that boiling at a high temperature is regarded as a particularly effective cooking method in the preparation of faunal and floral resources in pots [@Crown1995; @Stahl1989; @Wandsnider1997], strictly speaking, my study cannot reflect the entire subsistence change during the period in question. The most critical issue in here is whether there were more effective methods of cooking rice beside boiling in a pot. Rice is the center of existence in Asia, where more than 90 percent of the world's rice is grown. Traditionally, in East Asia (Korea, China, and Japan), the most familiar and well-known cooking method for rice is boiling with water [@Luh1980]. For special occasions, rice was used for making cakes, noodles or drinks [@Barker1985]. However, in terms of efficiency for the day to day consumption, boiling was the easiest way of cooking. Though we do not have solid evidence of how rice was cooked during the Mumun period, considering the known cooking methods, I argue that Mumun farmers probably preferred boiling. In addition, though it is reasonable to think that the pottery was mainly used as cooking vessels, rice might also have been cooked in bamboo tubes or other containers.

Second, the reference data used for CISA might not be suitable to the Korean Peninsula. Most of the reference data used in this thesis were generated by the modern wild fauna existing in Northern Europe [@Craig2011; @Copley2003; @Dudd1998; @Dudd1999; @Steele2010]. In order to avoid the effects of commercial farming and selective breeding, the modern reference samples were collected from authentic wild animals. The question is whether the $\delta$^13^C values of the available modern samples from Northern Europe are comparable with those of the archaeological ones from the Korean Peninsula. Generally, in Europe, there have been only rare examples of wild C~4~ plants [@Tafuri2009]. Therefore, wild herbivores mainly consumed C~3~ plants in general. In case of the Korean Peninsula, studies showed that most of indigenous wild plants are C~3~ plants; and the isotopic analysis revealed that the main food stuffs of wild animals are C~3~ plants [@Choy2010; @Ahn2006; @Kim2010; cf. @Lee2011a]. In these circumstances, one could argue that the basic environmental conditions which may affect the $\delta$^13^C values of the living organism in both regions are quite similar. However, I do admit that reference data used in this thesis might not perfectly reflect the environment in the Korean Peninsula. The best way to overcome this problem is to create CSIA reference data based on the indigenous fauna and flora from the Korean Peninsula. Future direction will focus on generating this reference.

Third, a pot is reused over time, and may be used to cook different kinds of food from one cooking episode to others. Since fatty acids and other compounds tend to accumulate in the fabric of the pot wall, the result of the organic geochemical analysis is more likely to reflect the entire usage of the pot. Also, because of the inherent nature of the isotope analysis, the result of CSIA is assumed to represent the type of food groups which were most frequently processed in it. This means that, even though the $\delta$^13^C values of a certain sample indicated the presence of porcine fat, and though I argue that the pot was mainly used for cooking pork, it is still possible that it was also used for cooking other food stuffs such as rice. In this regard, I also have to admit that the quantitative proportion of different food classes, which is derived from the analyses, cannot be assumed to be the direct representation of the diet taken by the ancient farmers of the site in question. Therefore, it is reasonable that the results of this thesis are viewed as a macroscopic explanation. Nevertheless, I think they have enough quality to give an insight into the human subsistence of the ancient Korea and the role of the intensive rice agriculture in the prehistoric Korean diet.

Fourth, it is possible that a certain type of pottery (or certain pot) was used for a certain type of food. As I mentioned in Chapter two, there are some variations in patterns on the Mumun potteries (Figure \label{Mpatterns}). In this setting, a pottery with a certain pattern might have been used for cooking rice. It is also possible that the frequency of usage might be different from one pot to another. For example, what if people ate rice for twenty meals a week, and other food for one meal a week, cooking the one and the other respectively in the same pot and in a number of different pots? In this case, linking the numbers of sampled sherds directly to the overall diet might not be proper for understanding the true nature of the subsistence. One way to overcome this limitation is to analyze sherds with different patterns and see if we can observe any trend. We can also investigate whether the frequency of usage can be distinguished by the organic geochemical analysis through the experiment with laboratory cooking episodes (e.g. one time vs. 10 times or more). Future researches will include these approaches.

Fifth, the study does not include the subsistence pattern of the Late Mumun period. Though not many studies have been conducted in relation to the subsistence of the Late Mumun period, it is assumed that rice was the main food stuff [cf. @Lee2011a]. To make a more convincing argument about the overall subsistence of the Mumun period, for future researches, I need to conduct the organic geochemical analysis on potsherds from the Late Mumun period village sites to see if I can observe a dramatic increase in rice consumption. Also, more Iron Age sites have to be included in the future researches. As I mentioned above, according to the paleobotanical evidence given by various places in the central part of the Korean Peninsula [@Lee2010; @Jeong2010], the Iron Age people had a diet focused on C~4~ plants. They mainly consumed different kinds of millets (foxtail, broomcorn and Japanese millet). On the other hand, the isotopic evidence from shell middens indicates that they consumed C~3~ plants, terrestrial animals and possibly marine resources [@Choy2009]. Though one Iron Age site, Eupha-Ri, was included in this thesis, the result was not enough to properly address the subsistence pattern during the period. Therefore, for better understanding, it is critical to analyze more potsherds from Iron Age villages.

Sixth, the study might have overlooked the role of domesticated animals in the ancient Korean farmers’ diet. In many cases, the places where we can observe the evidence of domesticated plants also tend to show that of domesticated animals. This is because the harvested crops from agriculture can be easily used for provisioning the livestock. Since domesticated plants may show a carbon isotope signal different from that of available indigenous wild plants, there is a strong possibility that domesticated animals show different $\delta$^13^C values, compared to the wild ones. The reference $\delta$^13^C value ranges I employed in this thesis were based on the data from wild animals, assuming the role of domesticated animals in the ancient Korean farmer’s diet is minimal at least until the Mumun period. Unfortunately, the domestication of animals is one of the least-studied areas in the Korean archaeology, mainly due to the high acidity of the sediments which does not allow long-term preservation of organisms. Among the available terrestrial mammals considered as major subsistence resources in the prehistoric Korean Peninsula, the strongest candidate for domestication is the pig. Though cattle, horse and dog were also considered, the main purpose of their domestication was not consumption. According to recent carbon and nitrogen isotope analyses on pig bones excavated from shell middens of the Korean Peninsula [@Lee2011], the isotopic signal of the pig shows it become omnivorous from the historic proto Kingdom period (ca. AD 0 - 250). Wild pigs are herbivores in general [@Lee2011], and both the $\delta$^13^C and $\delta$^15^N signal of pig bones from the Chulmun and Mumun periods were quite low (Figure \ref{bulk1}; \ref{bulk2}), indicating they mainly consumed wild C3 plants. These results support my assumption about the domesticated animals in the ancient Korean farmers’ diet. Overall, I think domesticated animals may have played a little role in subsistence throughout the prehistoric periods in Korea.

Lastly, the sampling of potsherds might have distorted the true nature of the subsistence pattern. Ideally, the entire pottery from each site need to be analyzed. However, due to the restriction in funding, sampling was inevitable. On one hand, since I wanted to maximize the representation of the entire site, I had to include the entire house pits that yield potteries. On the other hand, I had to compromise with myself in fixing the number of samples per one house pit, for the budget for the analyses was limited. In this regard, I admit that my sampling strategy might have distorted the real picture (for example, it is possible that we did not observe strong evidence of rice simply because the pots used to cook rice were not sampled). However, at the same time, I also think that this possibility is quite low, for I was able to observe a similar subsistence pattern at all the four sites included in this thesis.

\section{Summary}

I this discussion chapter, I re-evaluated the current rice-based model and its assumption of the strict dichotomy between Chulmun hunter-gatherers and Mumun full-dress rice farmers. Then, by correlating the results of the organic geochemical analyses and luminescence dating with available evidence, I made an argument about the role of rice as a subsistence strategy in the prehistoric Korean Peninsula. Lastly, I listed some of the important implications of the results in this thesis.

SeungkiKwak/Kwak_S_PhD_thesis documentation built on May 9, 2019, 1:22 p.m.