Researchers have debated on the extent of universality of classification systems in ethnobiology, with little attention to the fundamental mechanisms that underpin aboriginal taxonomy. In a mechanistic approach we challenge previous notions that aboriginal taxonomy is based predominantly on one particular mechanism (e.g., universal taxonomy, utilitarian, morphological, ecological). We hypothesize that traditional knowledge has evolved to include multifarious mechanisms, which provide a robust systematic classification and diverse taxonomy for plants. Quantitative classification techniques using Bray-Curtis average linkage and Detrended Correspondence Analysis (DCA) were employed to classify 50 plant species using both traditional aboriginal and Linnaean characters. Our research indicates that the Irulas of Tamil Nadu are an example of a traditional culture that uses three primary taxonomic mechanisms to identify plants; morphology, ecology and experience (“utility” was a secondary mechanism). Our ordination analysis suggests that the Irulas classification is more robust than the Linnaean taxonomy for the same group of plants.
Introduction
In the late 1960s, the study of folk classifications and folk taxonomy became vogue and academics began to focus on either cultural universals or the anthropological idea of cultural relativism. Folk classification is defined by the way in which members of a language community name and categorize plants and animals, whereas folk taxonomy is a term referring to the hierarchical structure, organic content, and cultural function of biological classifications that ethnobiologists find in every society around the world (Atran et al. 2004; Brown 2000). This early research in aboriginal classification was heavily influenced by linguistics.
Berlin and his colleagues (cf. Berlin et al. 1973) have pioneered the idea of universal categories in folk classification based on fundamentals in linguistics. They attest that, in contrast to social constructions of beauty, ritual, and social organization, human beings cannot construct the order in nature, but can only discern it (Berlin 1992). This means that many people in different cultures observe plants in basically the same way. They concluded that there are five, or possibly six, universal taxonomic ethnobiological categories in folk biology, which are arranged hierarchically and contain mutually exclusive taxa. These ethnobiological categories or ranks are similar to taxonomic ranks in Western zoology and botany (Berlin 1992) and have been used extensively by authors studying folk classification. Berlin and his colleagues argue that the categorization of taxa is based primarily on observed morphological (i.e., with a single mechanism) and behavioral affinities and differences among the recognized taxa. However, some researchers have debated on the extent of universality of classification systems in ethnobiology (Atran 1999; Ellen 1993) and mechanisms for aboriginal taxonomy are beginning to form in a variety of disciplines, including those within the natural sciences (Newmaster et al. 2006a).
Variation in morphology has been recognized as a principle mechanism in the aboriginal taxonomy of several biological organisms. In Posey's (1984) work with the Kayapo in the Brazilian Amazon, the most salient taxonomic units are those that reflect the natural variation in morphological forms. Costa-Neto (1998) also reports that the Pankarare Indians living in a semi-arid area of Brazil classify bees and wasps into precise ethnotaxonomical folk generic names based on morphology (Costa-Neto 1998). Although there are many reports of morphological mechanisms (Newmaster et al. 2006a), we found few studies that have recognized it exclusively in aboriginal taxonomy.
The utilitarian model is supported by considerable evidence as a mechanism in aboriginal taxonomy. Different cultures do exhibit differences in which plants and animals are recognized and named based on their usage. Ellen (1993) and Morris (2000) argue that the cultural importance of organisms is not accounted for adequately in Berlin's universal categories, and that folk classification systems will reflect the use and value placed on plants and animals which differ across cultures. The academics that hold more closely to cultural relativism in ethnobiological classification see folk science as an applied science that is rarely truly theoretical (Hayes 1982; Hunn 1982). Folk biological classification is seen as a special purpose classification that is driven by utilitarian (Hunn 1982), and social (Ellen 1993) concerns. Morris (2000) argues that ethnoscientists have underestimated the relevance of practical interests in the structuring of folk taxonomies, for utility is a major factor in classification of plants (Morris 2000). In his work with the Chewa people of Malawi, Morris notes that utility is of great importance in Chewa classifications, though he agrees that these people do have an interest in plant morphology and structure. Instead, he states that this interest in morphology is focused on the parts of the plants that are utilized, such as the leaves and roots for medicines (Morris 2000). The botanical taxonomies of the Lac Seul Anishinnaabe (Kenny and Parker 2004) or the Haida, Bella Coola, and Lillooet (Turner 1987) of Canada also resemble the classification systems of other foraging people in that some taxa have been represented based on cultural utility. Clement (1995) emphasizes that the intellectual and utilitarian values are the main purpose of classification, agreeing with the concepts of Hunn (1982), Morris (1984) and Brown (1976). Further, Clement (1995) concluded that traditional/aboriginal societies are not the only ones to use taxonomic features such as utilization values for grouping plants, but Linnaean taxonomy also incorporated a direct relationship between utility and taxonomic development.
A more recent mechanism in aboriginal taxonomy is ecology, which appears to also play an important role in how people classify flora and fauna of a given area. Hunn and French (1999) suggest that the more abundant and widely distributed an organism will be, the more likely it is to be noticed. Diamond and Bishop (1999) documented that the Ketengban (New Guinea) classified three species of cuckoos by habitat and altitudinal range. In India, the Kattunaikka classify plants using distinct ecosystems: the interior evergreen and moist deciduous forests, dry deciduous forests, rocky grasslands and marshy areas (Balakrishnan et al. 2003). Folk plant classification by communities around Kibale National Park in western Uganda divides green plants into major life form categories according to ecological characteristics such as habitat and dispersal mechanisms (Kakudidi 2004). Although there are few examples in the literature, it is likely that many more examples of ecological classifications will be discovered as this taxonomic mechanism is further investigated.
The scientific community and consequently modern Western civilization perceive plants using a Linnaean taxonomy, and largely ignores aboriginal taxonomy. We name, utilize and try to preserve the approximate 270,000 species of plants currently recognized on the globe (McKinney and Schoch 2003), which represents approximately 60% of the total plant species on earth (Bramwell 2002). We rarely consider aboriginal classifications in the context of ecological conservation and development, which has stimulated great concern about the increasing loss of this traditional aboriginal knowledge (Etkin 2002; Ohmagari and Berkes 1997).
Kanglin et al. (2000) found some important benefits of comparing scientific taxonomy with aboriginal classifications. Aboriginal classification is frequently quicker to use and simpler than the modern scientific taxonomy and is based on direct observation and macro-characters, while the Linnaean system often requires systematic research using microscopic morphology, molecular or even chemical characters. Diamond and Bishop (1999) write that the Ketengban people of New Guinea identify birds by song, silhouette, posture, and general appearance, without concern of fine details that Western scientists would be concerned about. Aboriginal classification can offer important clues with respect to the exploitation of the resources by local people, and thus is useful for commercial and government planning, and for scientific researchers. A more holistic view of plants must consider the taxonomic mechanisms used by aboriginal cultures in contrast with that of our modern Linnaean system, which is largely based on morphology (i.e., single mechanism).
We challenge the hypothesis that aboriginal taxonomy is based predominantly on one particular mechanism (e.g., universal taxonomy, utilitarian, morphological, ecological). We hypothesize that traditional knowledge has evolved to include multifarious mechanisms, which provide a robust systematic classification and diverse taxonomy for plants. In this study we explore the mechanism(s) responsible for classifying plants used by the ancient Irula culture of India using multivariate analyses. We will also compare Irula classification with that of the Linnaean classification system for the same group of plants. Little is known about the plant classification system of the Irulas, known as the “Forest People” in the remote areas of southern India, who have retained much of their traditional knowledge and way of life but are rapidly being mainstreamed into both regional as well as global culture.
Study Site
Kodiakkarai Reserve Forest (KRF) is a dry evergreen tropical forest located on the Coromandel Coast of Thanjavur district, Tamil Nadu, south India (Figure 1), 325 km south of Chennai, the capital of Tamil Nadu. It covers an area of about 25 km2 (Ragupathy 1992). Geographically it lies at latitude 10° 18′ N and longitude 79° 51′ E. The KRF is bordered by the Bay of Bengal on the east and Palk Strait on the south. The elevation ranges 25 to 50 m above the mean sea level. Like most other costal areas in Thanjavur District, the climate of KRF is fairly warm. The mean climate variables for our study area include: temperature 28°C, humidity 70%, and rainfall 1300 mm per year (Balasubramanian and Bole 1993; Parthasarathy and Karthikeyan 1997). The South West monsoon starts in June and continues till September.
Figure 1
Kodiakkarai Reserve Forest (KRF) located on the Coromandel coast of Thanjavur district, Tamil Nadu, east coast of India (Note: 1 = wet forest; 2 = dry evergreen forest; 3 = grassland). Modified from Ragupathy 1992.
Biodiversity in the KRF is high. The diversity of wildlife is rich and includes many tropical species, including over 240 species of birds (Balasubramanian 1992). Floristic diversity is also high within three main floristic regions; dry evergreen forest, wet forests (mangroves) and grasslands. The dry evergreen forest is dominated by many trees (e.g., Ochna obtusata DC., Ixora pavetta Andrews), shrubs (e.g., Salacia chinensis L., Dodonaea angustifolia L.f.) and vines (e.g., Cissus vitiginea L., Abrus precatorius L.). Many of the trees provide ideal habitats for epiphytic orchids and semi-parasitic plants such as Cymbidium aloifolium (L.) Sw. (Blasco and Legris 1972). The grassland consists of thorny shrubs (e.g., Dichrostachys cinerea (L.) Wight & Arn. and large patches of grass (e.g., Apluda mutica L., Dinebra retroflexa (Vahl) Panz.). Many herbaceous plants (e.g., Cleome tenella L.f., C. aspera Koenig ex. DC.) are found during the rainy season. The wetlands or “mangrove swamps” consists of both evergreen (e.g., Avicennia marina (Forssk.) Vierh. and deciduous (e.g., Excoecaria agallocha L.) vegetation intermixed with lianas (e.g., Pentatropis capensis (L. f.) Bullock), many shrubs (e.g., Scaevola plumieri Vahl) and herbs (e.g., Sesuvium portulacastrum L.) (Blasco and Legris 1972).
Ethnography of Irulas
The Irulas are a small tribal community that is part of the Dravidian language group spoken in southeastern India. The origin of the word "Irula" is not clear. It may have been derived from the Tamil word, Irul implying the dark complexion of the Irulas often spotted by villagers as distant silhouettes in the forests, hence the colloquial name the “Forest People” or kattu vasi (Fuchs 1973). They are recognized as a Scheduled Tribe (i.e., a special legal status eligible for Government subsidy and reservation policies) by the Government of India (Jeyadev and Ragupathy 1962).
Most ethnic populations of southern India are linguistically Dravidian (Breeks 1873; Hockings 1989). The Irulas belong to the Negrito (or Negroid) race, one of the six main ethnic groups that add to the racial mosaic of India. Negroids from Africa were the most ancient people to have come to India (Deepa et al. 2002). Their structure of language and the tribal dialects are slightly different from mainstream Dravidians (Jebadhas and Noble 1989) in south India who speak Telugu, Malyalam and Tamil. The Irulas are the largest Tamil-speaking aboriginal tribe in Tamil Nadu.
Many Irulas have been forced out of forest conservation reserves (Forest Protection Bill 1976). Some of them found themselves to be virtually bonded laborers, forced to live in urban areas (Devasahayam 1990). These Irulas started moving to the neighboring villages in hope of rebuilding their lives (Sharma 1996). However, some of the Irulas never left the forest and consequently their isolation from society has preserved an ancient culture over the last thousand years.
The ancient ways of the Irulas can still be observed today. Their staple food consists of minor millets, grain legumes, and wild yams supplemented with rice. They do not practice agriculture and therefore wholly depend on forest products and wild animals such as wild boar, rabbit, rat, deer, feral cat or kattu punai (‘kattu’ forest and ‘punai’ cat) and fish for their staple meat. Irula settlements are located on the outskirts of the forests and consist of tiny scattered huts. The community is divided into several exogamous clans. Marriage between inhabitants of different hamlets is generally encouraged and patrilocal residence is commonplace. More recently, some Irulas go to local villages to trade or sell honey, beeswax, firewood, wild fruits, yams, wild Carissa, Manilkara and Solanum berries, and other forest products in return for rice, potato, eggplant, tomato, tea, cookies and spices. Common occupations of the Irulas include harvesting of natural forest products, tobacco crops and the legendary profession snake “charming.” They have distinct cultural and religious practices, worshipping the Goddess Kanniamma and appointing Pujari (priests) within one family.
Methods
Data Collection
Field research was conducted in the Kodiakkarai Reserve Forest every month (covering all seasons) between 1990 and 1994 and then one final visit in the spring of 2005. Interview1 and data gathering methods followed that of Schultes (1960 and 1962), Jain (1991) and Bernard (1994). Interviews with healers and other knowledgeable inhabitants2 were conducted in order to understand how Irulas identify and utilize plants. Bernard's (1994) methodology was used to identify if Irulas were recognized as knowledgeable about the medicinal plants. In addition, we consulted the village head, who is familiar with the area and individuals who are authorities in botanical knowledge (aboriginal botanical knowledge). In total, 46 individuals (23 men and 23 women) between the ages of 20 and 60 were interviewed for extended periods (three days) on a repeated basis. Interviews were conducted in the regional language, Tamil. We also recorded information on the local customs, habits and beliefs with the help of the headman.
A second set of random interviews with a different subset of people were used to verify (> 85%) information for each species of plant considered. Only data verified from the secondary interviews (100% match) were used in the quantitative multivariate analyses. The interviews consisted of two formats: 1) our team collected plants with the Irulas in various ecosystems while conducting the interview; all species were collected with the Irulas at least five times in different habitats, and 2) a collection of the specimens was displayed stimulating the interview. This included a loose-leaf notebook of color photos of local plants and plant parts such as berries, stems, petioles, or roots. The data were gathered in a series of structured, semi-structured and unstructured interviews regarding plant uses, identification, and nomenclature during several field trips. The interview protocols, data confirmation and field observation were all followed as suggested by Bernard (1994); Etkin (1993); and Pelto and Pelto (1990). To elucidate cultural domains and determine differences in knowledge or taxonomy among the Irula, we cross-checked with other Irula respondents by using various research protocols such as free lists and pile sorts (Weller and Romney 1988; Werner and Fenton 1973).
Confirmation of identity for spontaneously described plants was by reference to fresh plant material collected to confirm the postulated identifications, and to “case” specimens of known identity (Gamble and Fischer 1915–1936; Hooker 1894). The Linnaean identities were designated by comparing the local specimens to type specimens from herbaria, and by referring to recent taxonomic monographs and revisions. The botanical nomenclature followed that of Flora of Tamil Nadu, India Series I Analysis (Henry et al. 1987, 1989; Nair and Henry 1983). Live species were propagated and can be found in the E.K. Janakiammal Ethnobotanical Garden, Botany Field Research Laboratory, University of Madras, Maduravoyal, Chennai. Herbarium voucher/type specimens have been deposited in the herbaria of University of Madras, Chennai and the Botanical Survey of India, Southern Circle.
Retention of Traditional Knowledge
We were curious to know if the other people from villages surrounding the periphery of the Kodiakkarai Reserve Forest (KRF) had a similar knowledge of native plants as demonstrated by the more traditional Irulas we interviewed. We conducted informal interviews with local residents along the sides of the roads surrounding the KRF and with traditional Irula living within the KRF. The interview consisted of five questions for a set of native plants commonly used by the Irulas. Five of the plants were common forest species and five plants were common grassland plants of which a couple are used as crops by the local farmers. The questions were: 1) Do you know this plant? 2) Do you know its name? 3) Is it wild or cultivated by local people? 4) What is use of the plant? and 5) From whom did you learn about the use of this plant?
Multivariate Analyses
Bray-Curtis average linkage was used to classify 50 species using the seven categories of characteristics used by the Irulas to identify plants. The similarity matrix of species by characters was used to build a cluster diagram, which is useful in assessing the ability of the characters in classifying all 50 species. Euclidean average linkage provided an independent check of the species membership defined by the Bray-Curtis average linkage. In the second part of the analysis, ordinations were used to assess the robustness of the classification and identify the most important characters in the classification as is routinely performed by Linnaean taxonomists (Gamble and Fischer 1915–1936; Hooker 1894).
Detrended Correspondence Analysis (DCA; ter Braak 1998) was used to produce three ordinations for 50 species using 1) seven Irula characteristics (morphology, ecology, experience, and four utilities: nutritional, medicinal, technical and ritual/spiritual), 2) fourteen Linnaean characteristics (leaf arrangement, leaf complexity, flower symmetry, ovary position, calyx number, calyx type, corolla number, corolla type, stamen number, stamen type, carpel number, carpel type, habit and other), and 3) both the Irula and Linnaean characteristics together. Interset Pearson correlations were calculated between every characteristic and each of the first four axes in each ordination. As an independent check, the relationship of classification structure in the species data to the taxonomic characteristics was analyzed with non-metric multidimensional scaling (NMS; Kruskal 1964; McCune and Mefford 1997; Primer 2002). In NMS, the Bray-Curtis distance measure was used because of its robustness for both large and small scales on the axes (Minchin 1987). Data were standardized by species maxima and two-dimensional solutions were appropriately chosen based on plotting a measure of fit (“stress”) to the number of dimensions. Stress represents distortion in the data and a stress value over 0.15 is high enough that the results are invalidated (Primer 2002). One hundred iterations were used for each NMS run, using random start coordinates. The first two ordination axes were rotated to enhance interpretability with the different axes.
Results
Plant Identification
The primary characteristics used by the Irulas to identify plants fall into three categories: morphology, experience and ecology. Morphological characters are used most often to recognize plants of which vegetative features (42%) are more commonly used than floral features (20%) (Figure 2). These vegetative and floral characteristics are macro and include several salient features such as shape, color and size. Personal experience with plants accounts for a significant portion of the skills used by the Irulas to identify plants (Figure 2). These experiences are hierarchical in usage and respectively comprised of olfaction (smell), palpation (touch) and gustation (taste). Ecological knowledge, such as where a plant lives, is another important skill utilized by the Irulas to identify plants (Figure 2) and is perhaps limited to the geographic region in which the study is conducted. In this study, the Irulas recognized the three most common biogeographic regions: dry forest, wet forest and grassland.
Plant Classification
The Irulas use a multi-mechanistic approach to classifying plants. A hierarchical cluster analysis resulted in a cluster diagram with well-developed branches and clusters representing several different taxonomic mechanisms (Figure 3). This classification included plant “utility” by the Irulas, which is defined as the use (e.g., nutritional, medicinal, technical or ritual). The foremost branch in the cluster diagram is predicated on utility and splits plants used by the Irulas for nutritional or medicinal use from that of those used for technical or ritualistic usage (Figure 3). The branch of this later group is further divided into plants used for technical and ritualistic purposes. All of the 50 species are further classified using morphology, experience or ecology. Some of the branches end in multiple species of which are further defined by specific characteristics common to that branch (i.e., the five species on the final branch labelled floral would be further classified by flower color, shape, etc.). We can clearly reject the hypothesis that a single mechanism is used by the Irulas to classify plants.
Figure 3
Classification of 50 species using Irula taxonomic characters (morphology, ecology, experience and utility: nutritional, medicinal, technical or ritual). The percentage similarity is based on the Bray-Curtis similarity measure. Irula nomenclature is bold and corresponding scientific names are italicized in parenthesis (Note: Exp. = Experience; Eco. = Ecology; Veg. = Vegetative; Morph. = Morphology; Nut. = Nutritional; Med. = Medicinal; Epi. = Epiphytic; Grs. = Grassland; Dry. = Dry evergreen forest; Wet. = Wetland).
Three primary taxonomic mechanisms are used by the Irulas to classify plants. Detrended correspondence analyses (DCA) was used to ordinate 50 species of plants used by the Irulas. Two theoretical gradients (eigenvalues axis 1 = 0.416; axis 2 = 0.221; NMS stress 0.11), of considerable length, were useful for separating all species into three groups (Table 1; Figure 4) using seven Irulas taxonomic characteristics (Table 2). These axes were used to explain 28.2% of the variance along the first axis and 15% of the data along the second axis. Non-metric multidimensional scaling confirmed the grouping of the species in the ordination into three groups. The three groups in the ordination were defined by three taxonomic mechanisms; ecology, morphology and experience. The first axis was most strongly, negatively correlated (−0.755; p < 0.001) with “ecosystem” (Table 2). The second axis was most strongly, positively correlated (0.837; p < 0.001) with “morphology” and negatively correlated (−0.676; p < 0.001) with “experience” (Table 2).
Table 1
Summary of Detrended Correspondence Analysis (DCA) for the ordination of 50 species classified using Irula taxonomy, Linnaean taxonomy or a combination of both.
Figure 4
DCA ordination of 50 plant species classified using Irula taxonomic characters. Correlations with DCA axis 1 (X-axis) and DCA axis 2 (Y-axis) are presented in table 1. Axes are scaled in standard deviations (ter Braak 1998).
Table 2
Pearson correlation of the DCA species scores with the seven Irula taxonomic categories (30 characters) for all 50 species (significant correlations are in bold and ranked by significance (* p < 0.05; **p <0.01; ***p <0.001)). Plants may have been used for up to 4 different types of utilities (e.g., nutritional, medicinal, technical or ritual).
Irula taxonomy was more extensive at classifying the 50 species studied than modern Linnaean taxonomy. Fourteen morphological characters (leaf arrangement, leaf complexity, flower symmetry, ovary position, calyx number, calyx type, corolla number, corolla type, stamen number, stamen type, carpel number, carpel type, habit and other) were used in a DCA ordination of the same 50 species used above. Although all species were clearly separated, there was no grouping of species and the eigenvalues were very low (axis 1 = 0.062; axis 2 = 0.029) and an independent check with NMS indicated invalid results (stress 0.16). A DCA of both Linnaean and Irulas characteristics resulted in relatively low eigenvalues that were slightly higher than that of the Linnaean DCA (Table 1).
Retention of Traditional Knowledge
From our survey of other communities outside of the KRF we found that 16 (2%) locals could answer three out of the five questions. Only 7% of the people could name more than one plant from the forest. We learned that those who responded grouped the plants into two sets: plants growing in forests (wild specimens) and those growing in the fields. These people knew more about the domesticated crop plants but relatively less about the wild plants. Some elderly people knew the name and utility of plants but were very poor in etymology of the plant name. Rarely could they identify the origin of their plant knowledge, but when identified they said it was orally passed as narrative by their elders, or they were taught to use the plant by an elder. The results from this interview suggest that little traditional knowledge is retained by lay people in contrast with the rich knowledge of plants possessed by the traditional Irulas living within the KRF who could answer all of the questions correctly and with ease.
Discussion
Comprehending The Traditional Knowledge of The Irulas
The Irulas live with the plants and have an intimate knowledge of each plant from seed to seedling, juvenile and adult forms. They have considerable ecological knowledge that is used to distinguish plants that look similar but are found in different habitats. For example, chavukku (Casuarina equisetifolia L.) is found near sand dunes of coastal areas, and resembles kattu chavukku (Tamarix indica Willd), which grows among mangroves in the tidal flats. The Irulas separate these two species on habitat alone. Their taxonomic ability to recognize different species is on par with those of most skilled Linnaean taxonomists. This is best exemplified by their ability to recognize arasa (Thespesia populneoides (Roxb.) Kostel.) of the mangrove swamps and poovaarasa (Thespesia populnea (L.) Solander ex Correa), which is found in grasslands and urban areas. Roxburgh, (1832) originally named Hibiscus populneoides Roxb., (different genus) and later Pierre, (1879–1907) treated this species as Thespesia populnea var. populneoides (Roxb.) Pierre. This species was further split up into two species by Fosberg and Sachet (1972) based on different habitats and morphology: 1) Thespesia populneoides (Roxb.) Kostel. is bushy with green leaves and grows in wetlands; and 2) Thespesia populnea (L.) Solander ex Correa is tall and is found in terrestrial habitats, its leaves bronzed with broader leaves and very shallowly cordate to subtruncate. According to Irulas, arasa and poovaarasa are clearly distinguished as separate species and are recognized primarily by the difference in habitat. We are currently using molecular markers (Newmaster et al. 2006b) to identify differences in these populations in order to clarify the Linnaean taxonomy as directed by the Irula classification.
The intimate knowledge of plants by the Irulas is closely associated with utility and their culture. To the Irulas, all plants have use, some of which are common uses and others are more ritualistic. The latter are topics that we will not define out of respect for the Irulas' beliefs and religion. This knowledge was shared intermittingly with us and we felt that it was a more personal sharing of knowledge that was never meant to be communicated with society at large. A detailed description of conventional uses of plants by the Irulas can be found in the work of Ragupathy and Mahadevan (1991, 1996).
Several sub-classification categories are worthy of further discussion. The Irulas group food plants according to their ‘hot-cold’ theory. Pregnant women, for example, are advised not to consume the wild fruit mulli (Solanum anguivi Lam.) or wild vegetable parupu keerai (Peplidium maritimum (L.f.) Asch.) because according to their ethos, these are cold- and heat-inducing plants and cause colds, influenza, stomach aches and related diseases. The young shoots of injai (Phoenix sylvestris (L.) Roxb.), moongil (Bambusa arundinacea (Retz.) Willd) or panai (Borassus flabellifer L.) are hot plants and consequently should be avoided during pregnancy because they may cause an abortion of the foetus. Many claims of medicinal utility were also listed during interviews about specific plants. For instance, nursing mothers are given ripe fruits of irrumbili (Diospyros ferrea (Wild.) Bahk. var. buxifolia) for enhanced milk secretion. Traditional medicinal uses of plants are in danger of being lost. A poisonous plant, adathoda (Justicia adhatoda L.), was traditionally used to treat lung ailments and eye pain. Today, this plant is known by the urban Irulas as “not edible by goats,” relating to its utility as a natural fence for goats. In this case the knowledge of a particular plant has become modernized or perhaps distorted over time.
Will The Irula Knowledge Be Lost?
The existence of the traditional Irulas or “Forest People” is under extreme pressure from a developing world that does not coalesce with the traditional way of living. The Irula population in the KRF is approximately 300. However, there a fewer people living in the remote forest areas than at the turn of the 20th century. Recent globalization is increasing the economic interests in natural resources in the plain forest (KRF) and the mountain forests (Western Ghats, India, Southeast Asia) by local state government and multinational corporations. As a result of the latter, biomass-based fuel and afforestation projects have being planned through “Self Help Groups” (including the Irulas), thereby accelerating the development of commodities from natural resources (Caruso and Reddy 2005). For example, pongan (Pongamia pinnata (L.) Pierre) is now in great demand because it has been billed as a new and cheap alternative fuel. Several international companies are currently developing natural resources in the forests and both the adivasi (i.e., traditional forest dwelling) and the local communities are not in a position to respond, or comprehend the long-term implications. Lawrence et al. (2004) suggests that we strengthen the capacity of Irula communities in order to contend with modern economic, cultural and livelihood practices with interventions that promote biodiversity rather than mono-cropping. The State Government has responded with policies for the adivasi, which provide compensation such as agricultural land and housing. Consequently the Irulas of the KRF are divided into two communities: those that live in the forest or on the outskirts of the forest, and those that live on the agricultural land given by the government. Some of the Irulas have taken up agriculture such as cultivating tobacco, while others have sold parts of their agricultural land to the local people, which often results in them working as bonded laborers. The Irula culture is threatened and subsequently so is there traditional knowledge.
The traditional elderly Irulas shared their concern regarding the loss of plants. They claim that many of the plants in the coastal forest are not found there any more. While on our expeditions we searched for one of these plants, velavarai, the white flowers of which were once used by the Irulas to cure blindness in elders. The identity of the famed velavarai is unknown, but it is thought to be similar to avarai (Cassia auriculata L.), which is commonly found in the outskirts of the KRF. Avarai is not used to cure blindness, but does have medicinal and ritualistic utility. The leaves, yellow flowers, fruits, stems and roots of avarai are used to maintain hygiene of the skin and scalp. Every year, the first day of the month of crop harvest is celebrated as the thanks-giving day for the “bull.” Locals pick many wild avarai flower bunches with the leaves and make a garland to tie it around the wounded neck of the festive bull. Perhaps the similar looking rare velavarai was over-harvested. To this date, repeated explorations could not find velavarai, including several expeditions by the Botanical Survey of India (Ragupathy and Mahadevan 1991). No herbarium records can be found for the famed velavarai. It is assumed that the routine practice and exploitation of this plant might have put it to extinction.
Taxonomic Mechanisms in Aboriginal Taxonomy
Classification systems and taxonomies are part of indigenous local knowledge, and the mechanisms of classification and identification for plants and animals in indigenous communities are triggered by the perceived morphological, utilitarian/cultural, and ecological affinities between organisms. While morphology plays an important role in folk classification systems, it is not the only mechanism, as it is in Linnaean taxonomy. Taxonomies need to be viewed, as Hayes (1982) suggests, in relation to “biological discontinuities in nature, chance historical events, utilitarian human concerns, human cultural concerns in a broader sense, intellectual curiosity, and constraints deriving from nature of human perception and cognition.” In 1982, Hunn challenged ethnobiologists to study classification systems with a combination of “cognitive, linguistic, ecological and evolutionary theory to define a dynamic ethnoecology.” Clement (1995), underscored the need for research on plant products alongside the study of morphological or behavioral classification in order to discover the relationship between cognitive and utilitarian factors in folk classification.
Ethnobiology explores the link between disciplines and perspectives that have been isolated in the past. Hayes (1982) and Hunn (1982), along with other colleagues, claimed that there is a utilitarian/adaptationist perspective to consider when looking at folk taxonomies, which challenges Berlin's cognitive/universal principles. This alternative view suggests that the environment is a setting in which people must adapt, and that folk biological classifications are a way by which people “systematically organize, store, and retrieve environmental information, which will enable them to accomplish this adaptation” (Hayes 1982). Morris (2000) sees the term taxonomic hierarchies as a misnomer, and agrees with Friedberg (1979; quoted in Morris 2000) that in ‘ethnographic reality’, folk taxonomies appear to be classified more according to a “complex web of resemblances” rather than forming a neat hierarchy. Morris also suggests that, even though ethnobiologists can accept that there is no necessary one-to-one relationship between utility and nomenclature, it is nevertheless important to recognize the taxonomic ordering that is intrinsically linked to functional conditions (Morris 2000). Morris (2000) challenges the assumptions made by Berlin, which propose that folk taxonomies are motivated by ‘intellectualist’ rather than by ‘utilitarianist’ considerations. A broader perspective including other disciplines (e.g., ecology, biology, etc.) should provide a more comprehensive understanding of how aboriginals classify and identify plants.
It is clear to us that aboriginal taxonomy of the Irulas is driven by several mechanisms. The Irulas use three primary taxonomic mechanisms: morphology, ecology and experience. We see utility as a fourth mechanism or more appropriately defined as a subcategory used to classify plants. It is unlikely that the mechanisms used by all cultures are the same or the same number. A growing body of scientific literature is in support of this hypothesis. We found that ecology was a primary mechanism in the Irulas plant taxonomy. Ecological salience however, does not stand alone, for the distribution of human observers within a local habitat is culturally conditioned (Hunn and French 1999). This means that different people within a culture have different knowledge and use of different ecosystems. Not only are folk classification categories not mutually exclusive, but identification and classification are triggered by many aspects of both the culture and the land in which it is situated. A holistic ethnobiological classification system begins with living organisms that are embedded within a place and within the relationships established between people and the plants found in that place (Davidson-Hunt et al. 2005). This is true for different cultures within India. The Malayali (another group of Dravidian tribe similar to Irulas in India) tribal farmers classify landraces of millet into species and varieties using plant morphology morphological, gastronomic characteristics (digestibility) and functional characteristics such as time for maturation and ecological tolerance (Rengalakshmi 2005). The Kattunaikka (yet another group of Dravidian tribe similar to Irulas in India) are experts in identifying various Dioscorea, mostly by relying on cooking qualities, texture, color, growth habit and smell of tubers. They also take into consideration other morphological characteristics such as size, shape and number of tubers (Balakrishnan et al. 2003). The Kattunaikka use ten different character classes and 44 character states of tubers and recognize five different ethnotaxa within the scientific Dioscorea pentaphylla (Balakrishnan et al. 2003). This classification is so detailed that it allows the Kattunaikka to recognize more taxa than professional Linnaean taxonomists (Balakrishnan et al. 2003).
There appears to be support for a multi-mechanistic taxonomy in other cultures. As reported by Atran (1985, 1999) the Itzaj Maya (Guatemala) use morphological characteristics and other biological properties to categorize life forms and generic species. Though morphological characteristics seem to be the primary trait that distinguishes life forms, they are also defined by their ecological roles, and adaptations to a broad array of ecological conditions (Atran 1999). Folk specifics and varietals, on the other hand, are classified by their cultural importance. Other ethnoscientists have recognized the cultural significance in the differentiation of folk generics (Berlin 1973). Kanglin et al. (2000), report on the botanical classification systems of the indigenous communities in Xishuangbanna, China. These local people classify plants, which have economic and function values based on local language, production practices, social customs, legends, economic utilization, morphological characteristics, and growth habits (Kanglin et al. 2000). Turner (2000) showed that not all categories in the folk classification are defined solely by morphological features. The ‘berries/fruit’ categories of the Lillooet are an example of the utilitarian categories. Turner (2000) suggests that there is no clear-cut distinction between morphological and utilitarian features of plants.
We suggest that the experiential mechanism is underestimated in its usage for classifying and identifying plants. Traditional knowledge evolves and is gained through contact with a plant. Our definition of the experiential mechanism is that cultures use sensory perception gained by experience as a tool for plant identification. This is not a novel concept and was coined the “anthropology of senses” by a small sub-discipline of scholars studying cultural “sensotypes” (Wober 1966). Their work has revealed cultural differences in linguistics that reflect a range of sensory perception. Howes (1988) defines the detailed classification and interpretation of smells across different cultures. This use of sensory perception to identify and utilize plants has been recorded in many cultures (Brett 1994; Casagrande 2001; Heinrich 1998; Leonti et al. 2002; Shepard 2002). Gollin (2004) focused on chemosensory selection criteria for medicinal plants. Her work revealed that the chemistry of less obvious sensory attributes is an excellent way to study relationships in human cognition, communication and therapeutic outcome. The Irulas in our study frequently interact with or experience plants. They do not hesitate to smell, touch or even taste different parts of the plant. Gollin recognized (1997a, 1997b) that aboriginal plant collectors in Kalimantran often inspected a plant's odor or taste as a habitual assessment of one's surroundings. We have also witnessed Anishinaabe and Métis traditional healers in northern Ontario, Canada, use this skill when collecting plants. Perhaps the most difficult identification skill to understand is that the Irulas recognize a small number of plants without the use of any characters. This is not an unfamiliar concept to taxonomists and is often dubbed the “Gestalt” identification (Kirchner 2000); we know some things by the whole and not by its finer physical properties. In most cases the Irulas were taught these plants at an early age without any reference to recognition characters.
A robust classification should be able to recognize more entities than a weaker classification. Our ordination analysis suggests that the Irulas classification is more robust than the Linnaean taxonomy for the same group of species. We attribute this to the multi-mechanistic taxonomy of the Irulas as apposed to a single mechanism (morphology) used in the Linnaean taxonomy. The design of our study did not allow us to test the hypothesis that the Irulas recognize more species than Linnaean taxonomists within a particular geographic area. However, there is some logic and empirical evidence to suggest that this hypothesis is worthy of study. It is logical to assume that aboriginal taxonomy is more developed because it has evolved over millennia versus the recent evolution of the modern Linnaean taxonomy (Genera Plantarum 1707–1718). Conklin (1954) reported that the Hanunoo of the Mindoro Island of the Philippines could identify approximately 1,600 different varieties of plants, whereas a contemporary botanical survey had recorded only 1,200 species. The Linnaean taxonomy of the genus Dioscorea is considered to be problematic (Prain and Burkill 1936; Velayudhan et al. 1998) because of the variability in its morphological characters, especially of aerial parts, such as leaves and bulbils. However, the Kattunaikka (another aboriginal group of Dravidian tribe similar to Irulas in India) who depend on wild Dioscorea species for their food classify at least 21 distinct taxa from just three Linnaean binomials (D. pentaphylla L., D. belophylla Voigt, and D. hamiltonii Hook. f.) (Balakrishnan et al. 2003). Similarly, Rengalakshmi (2005) reported how traditional Malayali tribal farmers in India recognize 15 varieties of millets from the four Linnaean binomials (Panicum sumatrense Roth ex Roem. & Schult., Panicum miliaceum L., Paspalum scrobiculatum L., and Setaria italica (L.) Beauv). Further study is needed to see if all cultures have robust taxonomies and if these cultures recognize more species than contemporary botanists. Modern molecular tools have been embraced by many Linnaean taxonomists/systematists, providing a second mechanism to classify plants. These molecular characters have considerable value in building the phylogenies of the biological species. We are currently embracing molecular markers in order to test the congruence of molecular and aboriginal classifications for intraspecific groups of taxa containing considerable molecular diversity. As with Cox (2000), we believe that cladistic analysis might be a useful method for cross-cultural ethnobotanical comparisons.
We must retain a caveat that aboriginal taxonomies are localized, as is the associated traditional knowledge. The lack of universality in aboriginal nomenclature is apparent in contrast with the Linnaean system. Distinct cultures or those who are geographically separated can name the same plant differently. In fact there are many examples of botanical surveys that recognize multiple names for the same plant with similar utility among different cultures (Balasubramaniam and Murugesan 2004; Ragupathy and Mahadevan 1991, 1996; Rengalakshmi 2005). The application of aboriginal taxonomy should remain local and not conflict with the universality of Linnaean taxonomy, which remains to be an excellent classification system for studying biology. Ethnobiologists need to consider both taxonomies in order to institute a holistic and integrative research program.
Conclusions
The desire for taxonomy (from the Greek word taxis ‘to arrange’) is an innate quality in humans that is instrumental in forming the foundations of our knowledge. To name and classify an object allows us to learn about that object and communicate this knowledge to others. The study of the mechanisms underlying aboriginal taxonomy is at the foundations of traditional knowledge. Understanding these traditional taxonomic mechanisms has certain urgency to it; aboriginal taxonomy has transcended millennia, but has degraded considerably in the last century. We hypothesize that there are several mechanisms, namely morphology, ecology, experience and utility as a subcategory. Furthermore these ancient classifications are more robust than our modern scientific classification. We are currently testing our predictions that traditional aboriginal taxonomists discriminate more species within their domain than Linnaean taxonomists, and whether molecular evidence is congruent with either type of classification. Further research is needed to understand if our multi-mechanistic hypothesis is supported by other cultures in remote areas of the globe.
Acknowledgments
We thank the wonderful people (Irulas) of KRF and surrounding villages for their enthusiasm and untiring support. A special thank you to Dr. K. Thothathri, Emeritus Scientist (Retd. Director, Botanical Survey of India), CAS in Botany, for his encouragement, suggestions and comments on a previous versions of the manuscript. We express our deepest appreciation to Mr. Sundaram Field Assistant), Kodiakadu Irula settlement (KRF) and his family members. He was instrumental for introducing us to many Irula herbalists, Irula elders and remote communities. A research fellowship award from the Botanical Survey of India, Calcutta, supported Mr. Sundaram. We are grateful to the late Prof. A. Mahadevan, Director, Centre for Advanced Study in Botany, University of Madras for his inspiring guidance and help throughout the course of investigation. We acknowledge Prof. N. Anand, Director, CAS in Botany, University of Madras, Dr. N. Madhivanan, and Mr. G. Surendran for their help in a recent trip (spring 2005) to KRF. Thanks to Jana Janakiraman, Kirit Patel and Candice Newmaster for reviewing an earlier version of this manuscript. Finally, we would like to thank Ian Smith and members of the Floristic Diversity Research Group, Biodiversity Institute of Ontario for the encouragement and support. We also thank two anonymous reviewers for their useful suggestions, and thank Christopher A. Peterson and Carmen Navarro Fernndez for French and Spanish translations.
